Hugging Face's logo

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

App Files Community
textgeflecht commited on
Commit
fdf2fa8
·
verified ·
1 Parent(s): ecea808

Upload microsoft_phi-4_2025-05-28-13-42-11.json

Browse files

ik_llama.cpp with modified code to send data: [DONE] at the end of a stream. run as: ./llama-server -m ./phi-4-Q8_0.gguf -c 16384 -ngl 99 -np 8 --cont-batching --host 0.0.0.0 --port 8000
with 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_2025-05-28-13-42-11.json +296 -0
results/microsoft_phi-4_2025-05-28-13-42-11.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": "ik_llama.cpp -np 8, 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": 10,
42
+ "total_tokens": 2028,
43
+ "token_throughput_secs": 70.47125620736956,
44
+ "duration_ms": 28777,
45
+ "time_to_first_token_ms": {
46
+ "p50": 60.59,
47
+ "p60": 61.796,
48
+ "p70": 63.439,
49
+ "p80": 64.946,
50
+ "p90": 81.602,
51
+ "p95": 155.918,
52
+ "p99": 215.37,
53
+ "avg": 75.779
54
+ },
55
+ "inter_token_latency_ms": {
56
+ "p50": 14.033,
57
+ "p60": 14.194,
58
+ "p70": 14.398,
59
+ "p80": 14.584,
60
+ "p90": 14.584,
61
+ "p95": 14.585,
62
+ "p99": 14.586,
63
+ "avg": 13.878
64
+ },
65
+ "failed_requests": 0,
66
+ "successful_requests": 10,
67
+ "request_rate": 0.3474914014170097,
68
+ "total_tokens_sent": 2000,
69
+ "e2e_latency_ms": {
70
+ "p50": 2864.745,
71
+ "p60": 2908.546,
72
+ "p70": 2956.951,
73
+ "p80": 2988.578,
74
+ "p90": 3093.512,
75
+ "p95": 3147.24,
76
+ "p99": 3190.223,
77
+ "avg": 2877.561
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": 22268,
90
+ "token_throughput_secs": 188.2822912572045,
91
+ "duration_ms": 118269,
92
+ "time_to_first_token_ms": {
93
+ "p50": 77.545,
94
+ "p60": 80.119,
95
+ "p70": 82.873,
96
+ "p80": 84.894,
97
+ "p90": 86.369,
98
+ "p95": 88.841,
99
+ "p99": 90.539,
100
+ "avg": 78.252
101
+ },
102
+ "inter_token_latency_ms": {
103
+ "p50": 21.732,
104
+ "p60": 21.861,
105
+ "p70": 21.941,
106
+ "p80": 22.088,
107
+ "p90": 22.237,
108
+ "p95": 22.379,
109
+ "p99": 22.553,
110
+ "avg": 21.692
111
+ },
112
+ "failed_requests": 0,
113
+ "successful_requests": 115,
114
+ "request_rate": 0.9723578001876466,
115
+ "total_tokens_sent": 23000,
116
+ "e2e_latency_ms": {
117
+ "p50": 4378.041,
118
+ "p60": 4425.06,
119
+ "p70": 4485.123,
120
+ "p80": 4522.143,
121
+ "p90": 4622.866,
122
+ "p95": 4718.184,
123
+ "p99": 4806.536,
124
+ "avg": 4258.713
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": 193,
136
+ "total_tokens": 37058,
137
+ "token_throughput_secs": 310.9991755719263,
138
+ "duration_ms": 119157,
139
+ "time_to_first_token_ms": {
140
+ "p50": 82.59,
141
+ "p60": 84.095,
142
+ "p70": 86.489,
143
+ "p80": 88.756,
144
+ "p90": 91.487,
145
+ "p95": 93.974,
146
+ "p99": 102.261,
147
+ "avg": 82.342
148
+ },
149
+ "inter_token_latency_ms": {
150
+ "p50": 24.484,
151
+ "p60": 24.589,
152
+ "p70": 24.683,
153
+ "p80": 24.805,
154
+ "p90": 25.007,
155
+ "p95": 25.16,
156
+ "p99": 25.839,
157
+ "avg": 24.636
158
+ },
159
+ "failed_requests": 0,
160
+ "successful_requests": 193,
161
+ "request_rate": 1.6196999537314958,
162
+ "total_tokens_sent": 38600,
163
+ "e2e_latency_ms": {
164
+ "p50": 4935.02,
165
+ "p60": 5008.802,
166
+ "p70": 5069.111,
167
+ "p80": 5184.869,
168
+ "p90": 5294.698,
169
+ "p95": 5378.949,
170
+ "p99": 5423.959,
171
+ "avg": 4767.408
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": 194,
183
+ "total_tokens": 37598,
184
+ "token_throughput_secs": 314.1879498529714,
185
+ "duration_ms": 119667,
186
+ "time_to_first_token_ms": {
187
+ "p50": 84.269,
188
+ "p60": 86.906,
189
+ "p70": 89.633,
190
+ "p80": 93.125,
191
+ "p90": 121.593,
192
+ "p95": 131.792,
193
+ "p99": 270.457,
194
+ "avg": 93.428
195
+ },
196
+ "inter_token_latency_ms": {
197
+ "p50": 24.607,
198
+ "p60": 24.689,
199
+ "p70": 24.771,
200
+ "p80": 24.901,
201
+ "p90": 25.13,
202
+ "p95": 25.295,
203
+ "p99": 25.736,
204
+ "avg": 24.62
205
+ },
206
+ "failed_requests": 0,
207
+ "successful_requests": 194,
208
+ "request_rate": 1.6211623562816229,
209
+ "total_tokens_sent": 38800,
210
+ "e2e_latency_ms": {
211
+ "p50": 4978.537,
212
+ "p60": 5044.063,
213
+ "p70": 5127.45,
214
+ "p80": 5227.95,
215
+ "p90": 5313.673,
216
+ "p95": 5443.518,
217
+ "p99": 5532.21,
218
+ "avg": 4841.162
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": 194,
230
+ "total_tokens": 37115,
231
+ "token_throughput_secs": 312.6109202546532,
232
+ "duration_ms": 118725,
233
+ "time_to_first_token_ms": {
234
+ "p50": 88.263,
235
+ "p60": 90.203,
236
+ "p70": 92.039,
237
+ "p80": 96.798,
238
+ "p90": 134.608,
239
+ "p95": 141.593,
240
+ "p99": 352.537,
241
+ "avg": 103.087
242
+ },
243
+ "inter_token_latency_ms": {
244
+ "p50": 24.765,
245
+ "p60": 24.961,
246
+ "p70": 25.076,
247
+ "p80": 25.221,
248
+ "p90": 25.462,
249
+ "p95": 25.679,
250
+ "p99": 27.503,
251
+ "avg": 24.788
252
+ },
253
+ "failed_requests": 0,
254
+ "successful_requests": 194,
255
+ "request_rate": 1.6340163957807547,
256
+ "total_tokens_sent": 38800,
257
+ "e2e_latency_ms": {
258
+ "p50": 4993.178,
259
+ "p60": 5051.48,
260
+ "p70": 5112.186,
261
+ "p80": 5219.659,
262
+ "p90": 5341.693,
263
+ "p95": 5428.007,
264
+ "p99": 5525.847,
265
+ "avg": 4817.913
266
+ }
267
+ }
268
+ ],
269
+ "start_time": "2025-05-28T13:34:12.239832087+00:00",
270
+ "end_time": "2025-05-28T13:42:11.658461959+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
+ }