Hugging Face's logo

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

App Files Community
textgeflecht commited on
Commit
8fd7140
·
verified ·
1 Parent(s): 74ef1b0

Upload microsoft_phi-4_2025-05-28-23-51-04.json

Browse files

llama cpp server: ./llama-server -m ./phi-4-Q8_0.gguf -c 65536 -ngl 99 -np 64 -fa --cont-batching --host 0.0.0.0 --port 8000
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 800 --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_2025-05-28-23-51-04.json +296 -0
results/microsoft_phi-4_2025-05-28-23-51-04.json ADDED
@@ -0,0 +1,296 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "max_vus": 800,
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 64 -fa -c 65536: 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": 13,
42
+ "total_tokens": 2633,
43
+ "token_throughput_secs": 83.69792771572823,
44
+ "duration_ms": 31458,
45
+ "time_to_first_token_ms": {
46
+ "p50": 44.725,
47
+ "p60": 45.086,
48
+ "p70": 45.297,
49
+ "p80": 45.414,
50
+ "p90": 45.427,
51
+ "p95": 112.018,
52
+ "p99": 191.927,
53
+ "avg": 57.079
54
+ },
55
+ "inter_token_latency_ms": {
56
+ "p50": 11.678,
57
+ "p60": 11.731,
58
+ "p70": 11.774,
59
+ "p80": 11.864,
60
+ "p90": 11.928,
61
+ "p95": 11.969,
62
+ "p99": 12.014,
63
+ "avg": 11.719
64
+ },
65
+ "failed_requests": 0,
66
+ "successful_requests": 13,
67
+ "request_rate": 0.41324461082585145,
68
+ "total_tokens_sent": 2600,
69
+ "e2e_latency_ms": {
70
+ "p50": 2415.345,
71
+ "p60": 2466.184,
72
+ "p70": 2504.008,
73
+ "p80": 2550.371,
74
+ "p90": 2571.383,
75
+ "p95": 2591.997,
76
+ "p99": 2612.54,
77
+ "avg": 2419.703
78
+ }
79
+ },
80
+ {
81
+ "id": "[email protected]/s",
82
+ "executor_type": "ConstantArrivalRate",
83
+ "config": {
84
+ "max_vus": 800,
85
+ "duration_secs": 120,
86
+ "rate": 1.0
87
+ },
88
+ "total_requests": 115,
89
+ "total_tokens": 22375,
90
+ "token_throughput_secs": 188.91317236669505,
91
+ "duration_ms": 118440,
92
+ "time_to_first_token_ms": {
93
+ "p50": 80.854,
94
+ "p60": 84.548,
95
+ "p70": 87.773,
96
+ "p80": 90.869,
97
+ "p90": 95.71,
98
+ "p95": 96.831,
99
+ "p99": 98.532,
100
+ "avg": 78.635
101
+ },
102
+ "inter_token_latency_ms": {
103
+ "p50": 20.812,
104
+ "p60": 20.985,
105
+ "p70": 21.124,
106
+ "p80": 21.26,
107
+ "p90": 21.43,
108
+ "p95": 21.536,
109
+ "p99": 21.605,
110
+ "avg": 20.328
111
+ },
112
+ "failed_requests": 0,
113
+ "successful_requests": 115,
114
+ "request_rate": 0.9709503831137399,
115
+ "total_tokens_sent": 23000,
116
+ "e2e_latency_ms": {
117
+ "p50": 4145.822,
118
+ "p60": 4200.826,
119
+ "p70": 4274.839,
120
+ "p80": 4363.682,
121
+ "p90": 4472.977,
122
+ "p95": 4593.863,
123
+ "p99": 4715.769,
124
+ "avg": 4018.416
125
+ }
126
+ },
127
+ {
128
+ "id": "[email protected]/s",
129
+ "executor_type": "ConstantArrivalRate",
130
+ "config": {
131
+ "max_vus": 800,
132
+ "duration_secs": 120,
133
+ "rate": 10.0
134
+ },
135
+ "total_requests": 372,
136
+ "total_tokens": 67709,
137
+ "token_throughput_secs": 586.9328880285384,
138
+ "duration_ms": 115360,
139
+ "time_to_first_token_ms": {
140
+ "p50": 19785.017,
141
+ "p60": 26640.891,
142
+ "p70": 33240.985,
143
+ "p80": 43284.45,
144
+ "p90": 49377.789,
145
+ "p95": 54399.816,
146
+ "p99": 59430.99,
147
+ "avg": 22656.022
148
+ },
149
+ "inter_token_latency_ms": {
150
+ "p50": 101.008,
151
+ "p60": 105.676,
152
+ "p70": 109.322,
153
+ "p80": 112.962,
154
+ "p90": 116.994,
155
+ "p95": 120.267,
156
+ "p99": 191.474,
157
+ "avg": 101.021
158
+ },
159
+ "failed_requests": 7,
160
+ "successful_requests": 365,
161
+ "request_rate": 3.1639886001922424,
162
+ "total_tokens_sent": 73000,
163
+ "e2e_latency_ms": {
164
+ "p50": 38530.995,
165
+ "p60": 46962.635,
166
+ "p70": 54067.376,
167
+ "p80": 64930.886,
168
+ "p90": 71918.912,
169
+ "p95": 76953.723,
170
+ "p99": 81801.492,
171
+ "avg": 40731.29
172
+ }
173
+ },
174
+ {
175
+ "id": "[email protected]/s",
176
+ "executor_type": "ConstantArrivalRate",
177
+ "config": {
178
+ "max_vus": 800,
179
+ "duration_secs": 120,
180
+ "rate": 30.0
181
+ },
182
+ "total_requests": 352,
183
+ "total_tokens": 65582,
184
+ "token_throughput_secs": 554.3312676198847,
185
+ "duration_ms": 118308,
186
+ "time_to_first_token_ms": {
187
+ "p50": 33185.594,
188
+ "p60": 45488.525,
189
+ "p70": 52366.292,
190
+ "p80": 65623.911,
191
+ "p90": 73298.349,
192
+ "p95": 80138.394,
193
+ "p99": 92823.97,
194
+ "avg": 36365.098
195
+ },
196
+ "inter_token_latency_ms": {
197
+ "p50": 104.249,
198
+ "p60": 109.354,
199
+ "p70": 112.301,
200
+ "p80": 117.372,
201
+ "p90": 121.015,
202
+ "p95": 125.989,
203
+ "p99": 157.551,
204
+ "avg": 100.597
205
+ },
206
+ "failed_requests": 4,
207
+ "successful_requests": 348,
208
+ "request_rate": 2.9414668831648907,
209
+ "total_tokens_sent": 69600,
210
+ "e2e_latency_ms": {
211
+ "p50": 53528.197,
212
+ "p60": 66215.057,
213
+ "p70": 74166.8,
214
+ "p80": 86870.033,
215
+ "p90": 96009.365,
216
+ "p95": 100937.696,
217
+ "p99": 106577.38,
218
+ "avg": 55283.231
219
+ }
220
+ },
221
+ {
222
+ "id": "[email protected]/s",
223
+ "executor_type": "ConstantArrivalRate",
224
+ "config": {
225
+ "max_vus": 800,
226
+ "duration_secs": 120,
227
+ "rate": 100.0
228
+ },
229
+ "total_requests": 150,
230
+ "total_tokens": 26162,
231
+ "token_throughput_secs": 218.92393364428733,
232
+ "duration_ms": 119502,
233
+ "time_to_first_token_ms": {
234
+ "p50": 67686.982,
235
+ "p60": 74178.119,
236
+ "p70": 79286.75,
237
+ "p80": 84393.998,
238
+ "p90": 88949.103,
239
+ "p95": 94396.241,
240
+ "p99": 112781.842,
241
+ "avg": 68170.587
242
+ },
243
+ "inter_token_latency_ms": {
244
+ "p50": 132.052,
245
+ "p60": 133.745,
246
+ "p70": 134.584,
247
+ "p80": 135.558,
248
+ "p90": 138.185,
249
+ "p95": 141.215,
250
+ "p99": 203.927,
251
+ "avg": 133.051
252
+ },
253
+ "failed_requests": 8,
254
+ "successful_requests": 142,
255
+ "request_rate": 1.1882577240841221,
256
+ "total_tokens_sent": 28400,
257
+ "e2e_latency_ms": {
258
+ "p50": 92129.798,
259
+ "p60": 97819.496,
260
+ "p70": 102756.725,
261
+ "p80": 109253.44,
262
+ "p90": 114319.949,
263
+ "p95": 116281.242,
264
+ "p99": 118312.268,
265
+ "avg": 92426.451
266
+ }
267
+ }
268
+ ],
269
+ "start_time": "2025-05-28T23:36:50.258540432+00:00",
270
+ "end_time": "2025-05-28T23:51:04.021377439+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
+ }