Hugging Face's logo

Spaces:
textgeflecht
/
inference-benchmarking-results-phi4-200-tokens
Running

App Files Community
textgeflecht commited on
Commit
305ff56
·
verified ·
1 Parent(s): f48b23e

Upload microsoft_phi-4_2025-05-28-14-56-02.json

Browse files

llama.cpp with ./llama-server -m ./phi-4-Q8_0.gguf -c 16384 -ngl 99 -np 8 -fa --cont-batching --host 0.0.0.0 --port 8000
and benchmark sudo docker run --network host -e HF_TOKEN=$HF_TOKEN -v ~/inference-benchmarker-results:/opt/inference-benchmarker/results inference_benchmarker inference-benchmarker --no-console --prompt-options "num_tokens=200,max_tokens=220,min_tokens=180,variance=10" --decode-options "num_tokens=200,max_tokens=220,min_tokens=180,variance=10" --url $URL --rates 1.0 --rates 10.0 --rates 30.0 --rates 100.0 --max-vus 8 --duration 120s --warmup 30s --benchmark-kind rate --model-name "phi-4" --tokenizer-name "microsoft/phi-4"

Files changed (1) hide show
  1. results/microsoft_phi-4_2025-05-28-14-56-02.json +296 -0
results/microsoft_phi-4_2025-05-28-14-56-02.json ADDED
@@ -0,0 +1,296 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "max_vus": 8,
4
+ "duration_secs": 120,
5
+ "benchmark_kind": "Rate",
6
+ "warmup_duration_secs": 30,
7
+ "rates": [
8
+ 1.0,
9
+ 10.0,
10
+ 30.0,
11
+ 100.0
12
+ ],
13
+ "num_rates": 10,
14
+ "prompt_options": {
15
+ "num_tokens": 200,
16
+ "min_tokens": 180,
17
+ "max_tokens": 220,
18
+ "variance": 10
19
+ },
20
+ "decode_options": {
21
+ "num_tokens": 200,
22
+ "min_tokens": 180,
23
+ "max_tokens": 220,
24
+ "variance": 10
25
+ },
26
+ "tokenizer": "microsoft/phi-4",
27
+ "model_name": "phi-4",
28
+ "profile": null,
29
+ "meta": null,
30
+ "run_id": "llama.cpp -np 8 -fa, max-vus 8: unsloth/phi-4-GGUF:Q8_0 (200 tokens)"
31
+ },
32
+ "results": [
33
+ {
34
+ "id": "warmup",
35
+ "executor_type": "ConstantVUs",
36
+ "config": {
37
+ "max_vus": 1,
38
+ "duration_secs": 30,
39
+ "rate": null
40
+ },
41
+ "total_requests": 15,
42
+ "total_tokens": 2667,
43
+ "token_throughput_secs": 84.76640145566599,
44
+ "duration_ms": 31462,
45
+ "time_to_first_token_ms": {
46
+ "p50": 43.923,
47
+ "p60": 44.081,
48
+ "p70": 44.176,
49
+ "p80": 44.293,
50
+ "p90": 44.638,
51
+ "p95": 47.967,
52
+ "p99": 53.97,
53
+ "avg": 44.608
54
+ },
55
+ "inter_token_latency_ms": {
56
+ "p50": 11.597,
57
+ "p60": 11.6,
58
+ "p70": 11.607,
59
+ "p80": 11.611,
60
+ "p90": 11.684,
61
+ "p95": 11.801,
62
+ "p99": 11.937,
63
+ "avg": 11.616
64
+ },
65
+ "failed_requests": 0,
66
+ "successful_requests": 15,
67
+ "request_rate": 0.4767514142613385,
68
+ "total_tokens_sent": 3000,
69
+ "e2e_latency_ms": {
70
+ "p50": 2285.188,
71
+ "p60": 2330.993,
72
+ "p70": 2371.983,
73
+ "p80": 2394.02,
74
+ "p90": 2455.957,
75
+ "p95": 2512.497,
76
+ "p99": 2545.905,
77
+ "avg": 2097.327
78
+ }
79
+ },
80
+ {
81
+ "id": "[email protected]/s",
82
+ "executor_type": "ConstantArrivalRate",
83
+ "config": {
84
+ "max_vus": 8,
85
+ "duration_secs": 120,
86
+ "rate": 1.0
87
+ },
88
+ "total_requests": 115,
89
+ "total_tokens": 22038,
90
+ "token_throughput_secs": 186.89396363364736,
91
+ "duration_ms": 117917,
92
+ "time_to_first_token_ms": {
93
+ "p50": 53.047,
94
+ "p60": 54.173,
95
+ "p70": 55.047,
96
+ "p80": 56.79,
97
+ "p90": 59.835,
98
+ "p95": 61.337,
99
+ "p99": 77.85,
100
+ "avg": 53.133
101
+ },
102
+ "inter_token_latency_ms": {
103
+ "p50": 22.077,
104
+ "p60": 22.436,
105
+ "p70": 22.806,
106
+ "p80": 23.143,
107
+ "p90": 23.685,
108
+ "p95": 24.117,
109
+ "p99": 25.089,
110
+ "avg": 21.569
111
+ },
112
+ "failed_requests": 0,
113
+ "successful_requests": 115,
114
+ "request_rate": 0.9752611769611328,
115
+ "total_tokens_sent": 23000,
116
+ "e2e_latency_ms": {
117
+ "p50": 4421.43,
118
+ "p60": 4504.873,
119
+ "p70": 4595.396,
120
+ "p80": 4680.992,
121
+ "p90": 4857.984,
122
+ "p95": 4924.05,
123
+ "p99": 5364.183,
124
+ "avg": 4171.345
125
+ }
126
+ },
127
+ {
128
+ "id": "[email protected]/s",
129
+ "executor_type": "ConstantArrivalRate",
130
+ "config": {
131
+ "max_vus": 8,
132
+ "duration_secs": 120,
133
+ "rate": 10.0
134
+ },
135
+ "total_requests": 180,
136
+ "total_tokens": 34587,
137
+ "token_throughput_secs": 288.95492547346106,
138
+ "duration_ms": 119696,
139
+ "time_to_first_token_ms": {
140
+ "p50": 55.657,
141
+ "p60": 58.53,
142
+ "p70": 60.406,
143
+ "p80": 128.016,
144
+ "p90": 254.077,
145
+ "p95": 289.023,
146
+ "p99": 347.043,
147
+ "avg": 94.422
148
+ },
149
+ "inter_token_latency_ms": {
150
+ "p50": 26.374,
151
+ "p60": 26.731,
152
+ "p70": 26.977,
153
+ "p80": 27.485,
154
+ "p90": 28.52,
155
+ "p95": 29.28,
156
+ "p99": 39.624,
157
+ "avg": 26.626
158
+ },
159
+ "failed_requests": 0,
160
+ "successful_requests": 180,
161
+ "request_rate": 1.5037987274184808,
162
+ "total_tokens_sent": 36000,
163
+ "e2e_latency_ms": {
164
+ "p50": 5243.4,
165
+ "p60": 5386.565,
166
+ "p70": 5496.027,
167
+ "p80": 5640.532,
168
+ "p90": 5821.668,
169
+ "p95": 6017.061,
170
+ "p99": 6254.725,
171
+ "avg": 5119.224
172
+ }
173
+ },
174
+ {
175
+ "id": "[email protected]/s",
176
+ "executor_type": "ConstantArrivalRate",
177
+ "config": {
178
+ "max_vus": 8,
179
+ "duration_secs": 120,
180
+ "rate": 30.0
181
+ },
182
+ "total_requests": 187,
183
+ "total_tokens": 35529,
184
+ "token_throughput_secs": 296.77264922144457,
185
+ "duration_ms": 119717,
186
+ "time_to_first_token_ms": {
187
+ "p50": 55.16,
188
+ "p60": 56.989,
189
+ "p70": 58.522,
190
+ "p80": 60.533,
191
+ "p90": 237.011,
192
+ "p95": 285.83,
193
+ "p99": 375.7,
194
+ "avg": 83.101
195
+ },
196
+ "inter_token_latency_ms": {
197
+ "p50": 26.521,
198
+ "p60": 26.943,
199
+ "p70": 27.293,
200
+ "p80": 27.77,
201
+ "p90": 28.204,
202
+ "p95": 29.268,
203
+ "p99": 35.763,
204
+ "avg": 26.238
205
+ },
206
+ "failed_requests": 0,
207
+ "successful_requests": 187,
208
+ "request_rate": 1.5620052746885682,
209
+ "total_tokens_sent": 37400,
210
+ "e2e_latency_ms": {
211
+ "p50": 5256.842,
212
+ "p60": 5348.763,
213
+ "p70": 5462.46,
214
+ "p80": 5551.717,
215
+ "p90": 5706.412,
216
+ "p95": 5846.258,
217
+ "p99": 6168.585,
218
+ "avg": 5011.004
219
+ }
220
+ },
221
+ {
222
+ "id": "[email protected]/s",
223
+ "executor_type": "ConstantArrivalRate",
224
+ "config": {
225
+ "max_vus": 8,
226
+ "duration_secs": 120,
227
+ "rate": 100.0
228
+ },
229
+ "total_requests": 190,
230
+ "total_tokens": 34964,
231
+ "token_throughput_secs": 293.59681469038,
232
+ "duration_ms": 119088,
233
+ "time_to_first_token_ms": {
234
+ "p50": 57.862,
235
+ "p60": 58.392,
236
+ "p70": 59.341,
237
+ "p80": 60.488,
238
+ "p90": 80.92,
239
+ "p95": 196.41,
240
+ "p99": 339.147,
241
+ "avg": 74.175
242
+ },
243
+ "inter_token_latency_ms": {
244
+ "p50": 26.834,
245
+ "p60": 27.273,
246
+ "p70": 27.657,
247
+ "p80": 28.017,
248
+ "p90": 28.657,
249
+ "p95": 29.223,
250
+ "p99": 40.72,
251
+ "avg": 27.19
252
+ },
253
+ "failed_requests": 0,
254
+ "successful_requests": 190,
255
+ "request_rate": 1.5954523164160905,
256
+ "total_tokens_sent": 38000,
257
+ "e2e_latency_ms": {
258
+ "p50": 5312.776,
259
+ "p60": 5414.764,
260
+ "p70": 5527.086,
261
+ "p80": 5609.911,
262
+ "p90": 5782.67,
263
+ "p95": 5851.942,
264
+ "p99": 6130.977,
265
+ "avg": 4923.735
266
+ }
267
+ }
268
+ ],
269
+ "start_time": "2025-05-28T14:47:17.869328796+00:00",
270
+ "end_time": "2025-05-28T14:56:02.455208643+00:00",
271
+ "system": {
272
+ "cpu": [
273
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu0@4699MHz",
274
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu1@4699MHz",
275
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu2@4699MHz",
276
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu3@4699MHz",
277
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu4@4699MHz",
278
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu5@4699MHz",
279
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu6@4699MHz",
280
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu7@4699MHz",
281
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu8@4699MHz",
282
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu9@4699MHz",
283
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu10@4699MHz",
284
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu11@4699MHz",
285
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu12@4699MHz",
286
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu13@4699MHz",
287
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu14@4699MHz",
288
+ "AMD Ryzen 7 9800X3D 8-Core Processor cpu15@4699MHz"
289
+ ],
290
+ "memory": "83.47 GB",
291
+ "os_name": "Debian GNU/Linux",
292
+ "os_version": "11",
293
+ "kernel": "5.15.167.4-microsoft-standard-WSL2",
294
+ "hostname": "computer"
295
+ }
296
+ }