hf-transformers-bot commited on
Commit
daf8134
·
verified ·
1 Parent(s): 5eb1f2e

Upload benchmark results for run 18261446187

Browse files
2025-10-05/18261446187/benchmark_results/Llama-2-7b-hf/Llama-2-7b-hf_benchmark_20251005_164929.json ADDED
@@ -0,0 +1,1175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_name": "Llama-2-7b-hf",
3
+ "benchmark_scenarios": [
4
+ {
5
+ "scenario_name": "eager_eager_attn",
6
+ "metadata": {
7
+ "timestamp": "2025-10-05T16:43:29.598378",
8
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
9
+ "hardware_info": {
10
+ "gpu_name": "NVIDIA A10G",
11
+ "gpu_memory_total_mb": 23028,
12
+ "cpu_count": 16,
13
+ "memory_total_mb": 63607,
14
+ "python_version": "3.10.12",
15
+ "torch_version": "2.8.0+cu126",
16
+ "cuda_version": "12.6"
17
+ },
18
+ "config": {
19
+ "name": "eager",
20
+ "model_id": "meta-llama/Llama-2-7b-hf",
21
+ "variant": "eager",
22
+ "warmup_iterations": 3,
23
+ "measurement_iterations": 5,
24
+ "num_tokens_to_generate": 100,
25
+ "device": "cuda",
26
+ "torch_dtype": "float16",
27
+ "compile_mode": null,
28
+ "compile_options": {},
29
+ "use_cache": true,
30
+ "batch_size": 1,
31
+ "sequence_length": null,
32
+ "attn_implementation": "eager",
33
+ "sdpa_backend": null,
34
+ "custom_params": {}
35
+ }
36
+ },
37
+ "measurements": {
38
+ "latency_seconds": {
39
+ "name": "latency_seconds",
40
+ "measurements": [
41
+ 3.657777099609375,
42
+ 3.651159912109375,
43
+ 3.65107666015625,
44
+ 3.648025634765625,
45
+ 3.64308837890625
46
+ ],
47
+ "mean": 3.6502255371093755,
48
+ "median": 3.65107666015625,
49
+ "std": 0.0047833494710593834,
50
+ "min": 3.64308837890625,
51
+ "max": 3.657777099609375,
52
+ "p25": 3.648025634765625,
53
+ "p75": 3.651159912109375,
54
+ "p90": 3.655130224609375,
55
+ "p95": 3.656453662109375,
56
+ "p99": 3.6575124121093747,
57
+ "unit": "seconds"
58
+ },
59
+ "time_to_first_token_seconds": {
60
+ "name": "time_to_first_token_seconds",
61
+ "measurements": [
62
+ 0.03764828872680664,
63
+ 0.03718809509277344,
64
+ 0.03706201553344726,
65
+ 0.03731840133666992,
66
+ 0.036887168884277344
67
+ ],
68
+ "mean": 0.03722079391479492,
69
+ "median": 0.03718809509277344,
70
+ "std": 0.0002568518917332893,
71
+ "min": 0.036887168884277344,
72
+ "max": 0.03764828872680664,
73
+ "p25": 0.03706201553344726,
74
+ "p75": 0.03731840133666992,
75
+ "p90": 0.037516333770751954,
76
+ "p95": 0.037582311248779296,
77
+ "p99": 0.03763509323120117,
78
+ "unit": "seconds"
79
+ },
80
+ "tokens_per_second": {
81
+ "name": "tokens_per_second",
82
+ "measurements": [
83
+ 27.3390087139753,
84
+ 27.388556625071857,
85
+ 27.38918113971358,
86
+ 27.412088074985444,
87
+ 27.449238008884816
88
+ ],
89
+ "mean": 27.3956145125262,
90
+ "median": 27.38918113971358,
91
+ "std": 0.035895058469016417,
92
+ "min": 27.3390087139753,
93
+ "max": 27.449238008884816,
94
+ "p25": 27.388556625071857,
95
+ "p75": 27.412088074985444,
96
+ "p90": 27.434378035325068,
97
+ "p95": 27.44180802210494,
98
+ "p99": 27.44775201152884,
99
+ "unit": "tokens/sec"
100
+ },
101
+ "time_per_output_token_seconds": {
102
+ "name": "time_per_output_token_seconds",
103
+ "measurements": [
104
+ 0.036577770996093745,
105
+ 0.03651159912109375,
106
+ 0.0365107666015625,
107
+ 0.036480256347656254,
108
+ 0.0364308837890625
109
+ ],
110
+ "mean": 0.03650225537109375,
111
+ "median": 0.0365107666015625,
112
+ "std": 4.783349471059273e-05,
113
+ "min": 0.0364308837890625,
114
+ "max": 0.036577770996093745,
115
+ "p25": 0.036480256347656254,
116
+ "p75": 0.03651159912109375,
117
+ "p90": 0.03655130224609375,
118
+ "p95": 0.03656453662109375,
119
+ "p99": 0.03657512412109375,
120
+ "unit": "seconds/token"
121
+ }
122
+ },
123
+ "gpu_metrics": {
124
+ "gpu_utilization_mean": 90.82608695652173,
125
+ "gpu_utilization_max": 94,
126
+ "gpu_utilization_min": 88,
127
+ "gpu_memory_used_mean": 13253,
128
+ "gpu_memory_used_max": 13253,
129
+ "gpu_memory_used_min": 13253,
130
+ "sample_count": 92,
131
+ "gpu_monitoring_status": "success"
132
+ }
133
+ },
134
+ {
135
+ "scenario_name": "compiled_compile_max-autotune_eager_attn",
136
+ "metadata": {
137
+ "timestamp": "2025-10-05T16:44:04.804224",
138
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
139
+ "hardware_info": {
140
+ "gpu_name": "NVIDIA A10G",
141
+ "gpu_memory_total_mb": 23028,
142
+ "cpu_count": 16,
143
+ "memory_total_mb": 63607,
144
+ "python_version": "3.10.12",
145
+ "torch_version": "2.8.0+cu126",
146
+ "cuda_version": "12.6"
147
+ },
148
+ "config": {
149
+ "name": "compiled",
150
+ "model_id": "meta-llama/Llama-2-7b-hf",
151
+ "variant": "compiled",
152
+ "warmup_iterations": 3,
153
+ "measurement_iterations": 5,
154
+ "num_tokens_to_generate": 100,
155
+ "device": "cuda",
156
+ "torch_dtype": "float16",
157
+ "compile_mode": "max-autotune",
158
+ "compile_options": {},
159
+ "use_cache": true,
160
+ "batch_size": 1,
161
+ "sequence_length": null,
162
+ "attn_implementation": "eager",
163
+ "sdpa_backend": null,
164
+ "custom_params": {}
165
+ }
166
+ },
167
+ "measurements": {
168
+ "latency_seconds": {
169
+ "name": "latency_seconds",
170
+ "measurements": [
171
+ 7.763421875,
172
+ 7.77922802734375,
173
+ 7.8036376953125,
174
+ 7.834384765625,
175
+ 7.8790791015625
176
+ ],
177
+ "mean": 7.81195029296875,
178
+ "median": 7.8036376953125,
179
+ "std": 0.04124181325925854,
180
+ "min": 7.763421875,
181
+ "max": 7.8790791015625,
182
+ "p25": 7.77922802734375,
183
+ "p75": 7.834384765625,
184
+ "p90": 7.861201367187499,
185
+ "p95": 7.8701402343749995,
186
+ "p99": 7.877291328125,
187
+ "unit": "seconds"
188
+ },
189
+ "time_to_first_token_seconds": {
190
+ "name": "time_to_first_token_seconds",
191
+ "measurements": [
192
+ 0.038932960510253904,
193
+ 0.038012863159179684,
194
+ 0.03857875061035156,
195
+ 0.03809785461425781,
196
+ 0.03811475372314453
197
+ ],
198
+ "mean": 0.03834743652343751,
199
+ "median": 0.03811475372314453,
200
+ "std": 0.0003534731381463129,
201
+ "min": 0.038012863159179684,
202
+ "max": 0.038932960510253904,
203
+ "p25": 0.03809785461425781,
204
+ "p75": 0.03857875061035156,
205
+ "p90": 0.03879127655029297,
206
+ "p95": 0.03886211853027344,
207
+ "p99": 0.03891879211425781,
208
+ "unit": "seconds"
209
+ },
210
+ "tokens_per_second": {
211
+ "name": "tokens_per_second",
212
+ "measurements": [
213
+ 12.880917926413732,
214
+ 12.854745952747887,
215
+ 12.814536489830651,
216
+ 12.764244161043877,
217
+ 12.691838565266975
218
+ ],
219
+ "mean": 12.801256619060624,
220
+ "median": 12.814536489830651,
221
+ "std": 0.06741753513680784,
222
+ "min": 12.691838565266975,
223
+ "max": 12.880917926413732,
224
+ "p25": 12.764244161043877,
225
+ "p75": 12.854745952747887,
226
+ "p90": 12.870449136947395,
227
+ "p95": 12.875683531680563,
228
+ "p99": 12.879871047467098,
229
+ "unit": "tokens/sec"
230
+ },
231
+ "time_per_output_token_seconds": {
232
+ "name": "time_per_output_token_seconds",
233
+ "measurements": [
234
+ 0.07763421875,
235
+ 0.0777922802734375,
236
+ 0.078036376953125,
237
+ 0.07834384765624999,
238
+ 0.078790791015625
239
+ ],
240
+ "mean": 0.0781195029296875,
241
+ "median": 0.078036376953125,
242
+ "std": 0.000412418132592586,
243
+ "min": 0.07763421875,
244
+ "max": 0.078790791015625,
245
+ "p25": 0.0777922802734375,
246
+ "p75": 0.07834384765624999,
247
+ "p90": 0.078612013671875,
248
+ "p95": 0.07870140234375,
249
+ "p99": 0.07877291328125,
250
+ "unit": "seconds/token"
251
+ }
252
+ },
253
+ "gpu_metrics": {
254
+ "gpu_utilization_mean": 44.14705882352941,
255
+ "gpu_utilization_max": 92,
256
+ "gpu_utilization_min": 0,
257
+ "gpu_memory_used_mean": 13629.835294117647,
258
+ "gpu_memory_used_max": 13777,
259
+ "gpu_memory_used_min": 13415,
260
+ "sample_count": 170,
261
+ "gpu_monitoring_status": "success"
262
+ }
263
+ },
264
+ {
265
+ "scenario_name": "eager_sdpa_default",
266
+ "metadata": {
267
+ "timestamp": "2025-10-05T16:45:31.950259",
268
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
269
+ "hardware_info": {
270
+ "gpu_name": "NVIDIA A10G",
271
+ "gpu_memory_total_mb": 23028,
272
+ "cpu_count": 16,
273
+ "memory_total_mb": 63607,
274
+ "python_version": "3.10.12",
275
+ "torch_version": "2.8.0+cu126",
276
+ "cuda_version": "12.6"
277
+ },
278
+ "config": {
279
+ "name": "eager",
280
+ "model_id": "meta-llama/Llama-2-7b-hf",
281
+ "variant": "eager",
282
+ "warmup_iterations": 3,
283
+ "measurement_iterations": 5,
284
+ "num_tokens_to_generate": 100,
285
+ "device": "cuda",
286
+ "torch_dtype": "float16",
287
+ "compile_mode": null,
288
+ "compile_options": {},
289
+ "use_cache": true,
290
+ "batch_size": 1,
291
+ "sequence_length": null,
292
+ "attn_implementation": "sdpa",
293
+ "sdpa_backend": null,
294
+ "custom_params": {}
295
+ }
296
+ },
297
+ "measurements": {
298
+ "latency_seconds": {
299
+ "name": "latency_seconds",
300
+ "measurements": [
301
+ 3.4084306640625,
302
+ 3.408165771484375,
303
+ 3.4083876953125,
304
+ 3.409863037109375,
305
+ 3.4067138671875
306
+ ],
307
+ "mean": 3.40831220703125,
308
+ "median": 3.4083876953125,
309
+ "std": 0.0010000938061586206,
310
+ "min": 3.4067138671875,
311
+ "max": 3.409863037109375,
312
+ "p25": 3.408165771484375,
313
+ "p75": 3.4084306640625,
314
+ "p90": 3.4092900878906254,
315
+ "p95": 3.4095765625000003,
316
+ "p99": 3.4098057421875003,
317
+ "unit": "seconds"
318
+ },
319
+ "time_to_first_token_seconds": {
320
+ "name": "time_to_first_token_seconds",
321
+ "measurements": [
322
+ 0.03523977661132813,
323
+ 0.03461727905273437,
324
+ 0.034579326629638675,
325
+ 0.034541217803955075,
326
+ 0.034595455169677734
327
+ ],
328
+ "mean": 0.0347146110534668,
329
+ "median": 0.034595455169677734,
330
+ "std": 0.00026375655460241084,
331
+ "min": 0.034541217803955075,
332
+ "max": 0.03523977661132813,
333
+ "p25": 0.034579326629638675,
334
+ "p75": 0.03461727905273437,
335
+ "p90": 0.034990777587890626,
336
+ "p95": 0.03511527709960938,
337
+ "p99": 0.03521487670898438,
338
+ "unit": "seconds"
339
+ },
340
+ "tokens_per_second": {
341
+ "name": "tokens_per_second",
342
+ "measurements": [
343
+ 29.339015475471122,
344
+ 29.341295789273335,
345
+ 29.339385345607358,
346
+ 29.326691105098597,
347
+ 29.353800729545146
348
+ ],
349
+ "mean": 29.340037688999114,
350
+ "median": 29.339385345607358,
351
+ "std": 0.00860936554669196,
352
+ "min": 29.326691105098597,
353
+ "max": 29.353800729545146,
354
+ "p25": 29.339015475471122,
355
+ "p75": 29.341295789273335,
356
+ "p90": 29.348798753436423,
357
+ "p95": 29.351299741490784,
358
+ "p99": 29.353300531934273,
359
+ "unit": "tokens/sec"
360
+ },
361
+ "time_per_output_token_seconds": {
362
+ "name": "time_per_output_token_seconds",
363
+ "measurements": [
364
+ 0.034084306640625,
365
+ 0.03408165771484375,
366
+ 0.034083876953125,
367
+ 0.03409863037109375,
368
+ 0.034067138671875
369
+ ],
370
+ "mean": 0.0340831220703125,
371
+ "median": 0.034083876953125,
372
+ "std": 1.0000938061586512e-05,
373
+ "min": 0.034067138671875,
374
+ "max": 0.03409863037109375,
375
+ "p25": 0.03408165771484375,
376
+ "p75": 0.034084306640625,
377
+ "p90": 0.03409290087890625,
378
+ "p95": 0.034095765625,
379
+ "p99": 0.034098057421875,
380
+ "unit": "seconds/token"
381
+ }
382
+ },
383
+ "gpu_metrics": {
384
+ "gpu_utilization_mean": 96.96511627906976,
385
+ "gpu_utilization_max": 98,
386
+ "gpu_utilization_min": 95,
387
+ "gpu_memory_used_mean": 13849,
388
+ "gpu_memory_used_max": 13849,
389
+ "gpu_memory_used_min": 13849,
390
+ "sample_count": 86,
391
+ "gpu_monitoring_status": "success"
392
+ }
393
+ },
394
+ {
395
+ "scenario_name": "eager_sdpa_math",
396
+ "metadata": {
397
+ "timestamp": "2025-10-05T16:46:04.396986",
398
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
399
+ "hardware_info": {
400
+ "gpu_name": "NVIDIA A10G",
401
+ "gpu_memory_total_mb": 23028,
402
+ "cpu_count": 16,
403
+ "memory_total_mb": 63607,
404
+ "python_version": "3.10.12",
405
+ "torch_version": "2.8.0+cu126",
406
+ "cuda_version": "12.6"
407
+ },
408
+ "config": {
409
+ "name": "eager",
410
+ "model_id": "meta-llama/Llama-2-7b-hf",
411
+ "variant": "eager",
412
+ "warmup_iterations": 3,
413
+ "measurement_iterations": 5,
414
+ "num_tokens_to_generate": 100,
415
+ "device": "cuda",
416
+ "torch_dtype": "float16",
417
+ "compile_mode": null,
418
+ "compile_options": {},
419
+ "use_cache": true,
420
+ "batch_size": 1,
421
+ "sequence_length": null,
422
+ "attn_implementation": "sdpa",
423
+ "sdpa_backend": "math",
424
+ "custom_params": {}
425
+ }
426
+ },
427
+ "measurements": {
428
+ "latency_seconds": {
429
+ "name": "latency_seconds",
430
+ "measurements": [
431
+ 3.5345654296875,
432
+ 3.531876953125,
433
+ 3.53701025390625,
434
+ 3.52957666015625,
435
+ 3.5307724609375
436
+ ],
437
+ "mean": 3.5327603515625,
438
+ "median": 3.531876953125,
439
+ "std": 0.0026902776762012235,
440
+ "min": 3.52957666015625,
441
+ "max": 3.53701025390625,
442
+ "p25": 3.5307724609375,
443
+ "p75": 3.5345654296875,
444
+ "p90": 3.5360323242187497,
445
+ "p95": 3.5365212890624997,
446
+ "p99": 3.5369124609375,
447
+ "unit": "seconds"
448
+ },
449
+ "time_to_first_token_seconds": {
450
+ "name": "time_to_first_token_seconds",
451
+ "measurements": [
452
+ 0.03961916732788086,
453
+ 0.038910270690917965,
454
+ 0.038813278198242186,
455
+ 0.037330337524414066,
456
+ 0.0375052490234375
457
+ ],
458
+ "mean": 0.038435660552978514,
459
+ "median": 0.038813278198242186,
460
+ "std": 0.000878173418758148,
461
+ "min": 0.037330337524414066,
462
+ "max": 0.03961916732788086,
463
+ "p25": 0.0375052490234375,
464
+ "p75": 0.038910270690917965,
465
+ "p90": 0.0393356086730957,
466
+ "p95": 0.03947738800048828,
467
+ "p99": 0.03959081146240234,
468
+ "unit": "seconds"
469
+ },
470
+ "tokens_per_second": {
471
+ "name": "tokens_per_second",
472
+ "measurements": [
473
+ 28.29202118033539,
474
+ 28.313557161588015,
475
+ 28.272465393494603,
476
+ 28.332009651144144,
477
+ 28.3224141760321
478
+ ],
479
+ "mean": 28.306493512518852,
480
+ "median": 28.313557161588015,
481
+ "std": 0.021549001268627223,
482
+ "min": 28.272465393494603,
483
+ "max": 28.332009651144144,
484
+ "p25": 28.29202118033539,
485
+ "p75": 28.3224141760321,
486
+ "p90": 28.328171461099327,
487
+ "p95": 28.330090556121736,
488
+ "p99": 28.33162583213966,
489
+ "unit": "tokens/sec"
490
+ },
491
+ "time_per_output_token_seconds": {
492
+ "name": "time_per_output_token_seconds",
493
+ "measurements": [
494
+ 0.035345654296875,
495
+ 0.035318769531249995,
496
+ 0.035370102539062497,
497
+ 0.0352957666015625,
498
+ 0.035307724609375
499
+ ],
500
+ "mean": 0.035327603515625,
501
+ "median": 0.035318769531249995,
502
+ "std": 2.6902776762012287e-05,
503
+ "min": 0.0352957666015625,
504
+ "max": 0.035370102539062497,
505
+ "p25": 0.035307724609375,
506
+ "p75": 0.035345654296875,
507
+ "p90": 0.0353603232421875,
508
+ "p95": 0.035365212890625,
509
+ "p99": 0.035369124609375,
510
+ "unit": "seconds/token"
511
+ }
512
+ },
513
+ "gpu_metrics": {
514
+ "gpu_utilization_mean": 96.10112359550561,
515
+ "gpu_utilization_max": 98,
516
+ "gpu_utilization_min": 89,
517
+ "gpu_memory_used_mean": 13849,
518
+ "gpu_memory_used_max": 13849,
519
+ "gpu_memory_used_min": 13849,
520
+ "sample_count": 89,
521
+ "gpu_monitoring_status": "success"
522
+ }
523
+ },
524
+ {
525
+ "scenario_name": "eager_sdpa_flash_attention",
526
+ "metadata": {
527
+ "timestamp": "2025-10-05T16:46:37.840613",
528
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
529
+ "hardware_info": {
530
+ "gpu_name": "NVIDIA A10G",
531
+ "gpu_memory_total_mb": 23028,
532
+ "cpu_count": 16,
533
+ "memory_total_mb": 63607,
534
+ "python_version": "3.10.12",
535
+ "torch_version": "2.8.0+cu126",
536
+ "cuda_version": "12.6"
537
+ },
538
+ "config": {
539
+ "name": "eager",
540
+ "model_id": "meta-llama/Llama-2-7b-hf",
541
+ "variant": "eager",
542
+ "warmup_iterations": 3,
543
+ "measurement_iterations": 5,
544
+ "num_tokens_to_generate": 100,
545
+ "device": "cuda",
546
+ "torch_dtype": "float16",
547
+ "compile_mode": null,
548
+ "compile_options": {},
549
+ "use_cache": true,
550
+ "batch_size": 1,
551
+ "sequence_length": null,
552
+ "attn_implementation": "sdpa",
553
+ "sdpa_backend": "flash_attention",
554
+ "custom_params": {}
555
+ }
556
+ },
557
+ "measurements": {
558
+ "latency_seconds": {
559
+ "name": "latency_seconds",
560
+ "measurements": [
561
+ 3.408860595703125,
562
+ 3.411012939453125,
563
+ 3.4082529296875,
564
+ 3.40781201171875,
565
+ 3.409111572265625
566
+ ],
567
+ "mean": 3.4090100097656246,
568
+ "median": 3.408860595703125,
569
+ "std": 0.0011002487356555806,
570
+ "min": 3.40781201171875,
571
+ "max": 3.411012939453125,
572
+ "p25": 3.4082529296875,
573
+ "p75": 3.409111572265625,
574
+ "p90": 3.410252392578125,
575
+ "p95": 3.410632666015625,
576
+ "p99": 3.410936884765625,
577
+ "unit": "seconds"
578
+ },
579
+ "time_to_first_token_seconds": {
580
+ "name": "time_to_first_token_seconds",
581
+ "measurements": [
582
+ 0.035265632629394535,
583
+ 0.03474118423461914,
584
+ 0.03467993545532227,
585
+ 0.0346577262878418,
586
+ 0.03477932739257812
587
+ ],
588
+ "mean": 0.03482476119995117,
589
+ "median": 0.03474118423461914,
590
+ "std": 0.00022462952408651,
591
+ "min": 0.0346577262878418,
592
+ "max": 0.035265632629394535,
593
+ "p25": 0.03467993545532227,
594
+ "p75": 0.03477932739257812,
595
+ "p90": 0.03507111053466797,
596
+ "p95": 0.03516837158203125,
597
+ "p99": 0.03524618041992188,
598
+ "unit": "seconds"
599
+ },
600
+ "tokens_per_second": {
601
+ "name": "tokens_per_second",
602
+ "measurements": [
603
+ 29.335315185974512,
604
+ 29.316804648660355,
605
+ 29.340545453347243,
606
+ 29.344341664423098,
607
+ 29.33315553927209
608
+ ],
609
+ "mean": 29.33403249833546,
610
+ "median": 29.335315185974512,
611
+ "std": 0.009464783543369866,
612
+ "min": 29.316804648660355,
613
+ "max": 29.344341664423098,
614
+ "p25": 29.33315553927209,
615
+ "p75": 29.340545453347243,
616
+ "p90": 29.342823179992756,
617
+ "p95": 29.343582422207927,
618
+ "p99": 29.344189815980062,
619
+ "unit": "tokens/sec"
620
+ },
621
+ "time_per_output_token_seconds": {
622
+ "name": "time_per_output_token_seconds",
623
+ "measurements": [
624
+ 0.03408860595703125,
625
+ 0.03411012939453125,
626
+ 0.034082529296875,
627
+ 0.0340781201171875,
628
+ 0.034091115722656246
629
+ ],
630
+ "mean": 0.03409010009765624,
631
+ "median": 0.03408860595703125,
632
+ "std": 1.100248735655615e-05,
633
+ "min": 0.0340781201171875,
634
+ "max": 0.03411012939453125,
635
+ "p25": 0.034082529296875,
636
+ "p75": 0.034091115722656246,
637
+ "p90": 0.034102523925781246,
638
+ "p95": 0.03410632666015625,
639
+ "p99": 0.03410936884765625,
640
+ "unit": "seconds/token"
641
+ }
642
+ },
643
+ "gpu_metrics": {
644
+ "gpu_utilization_mean": 96.95348837209302,
645
+ "gpu_utilization_max": 98,
646
+ "gpu_utilization_min": 94,
647
+ "gpu_memory_used_mean": 13849,
648
+ "gpu_memory_used_max": 13849,
649
+ "gpu_memory_used_min": 13849,
650
+ "sample_count": 86,
651
+ "gpu_monitoring_status": "success"
652
+ }
653
+ },
654
+ {
655
+ "scenario_name": "eager_sdpa_efficient_attention",
656
+ "metadata": {
657
+ "timestamp": "2025-10-05T16:47:10.240016",
658
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
659
+ "hardware_info": {
660
+ "gpu_name": "NVIDIA A10G",
661
+ "gpu_memory_total_mb": 23028,
662
+ "cpu_count": 16,
663
+ "memory_total_mb": 63607,
664
+ "python_version": "3.10.12",
665
+ "torch_version": "2.8.0+cu126",
666
+ "cuda_version": "12.6"
667
+ },
668
+ "config": {
669
+ "name": "eager",
670
+ "model_id": "meta-llama/Llama-2-7b-hf",
671
+ "variant": "eager",
672
+ "warmup_iterations": 3,
673
+ "measurement_iterations": 5,
674
+ "num_tokens_to_generate": 100,
675
+ "device": "cuda",
676
+ "torch_dtype": "float16",
677
+ "compile_mode": null,
678
+ "compile_options": {},
679
+ "use_cache": true,
680
+ "batch_size": 1,
681
+ "sequence_length": null,
682
+ "attn_implementation": "sdpa",
683
+ "sdpa_backend": "efficient_attention",
684
+ "custom_params": {}
685
+ }
686
+ },
687
+ "measurements": {
688
+ "latency_seconds": {
689
+ "name": "latency_seconds",
690
+ "measurements": [
691
+ 3.409301025390625,
692
+ 3.40593603515625,
693
+ 3.403377197265625,
694
+ 3.408706298828125,
695
+ 3.4053486328125
696
+ ],
697
+ "mean": 3.4065338378906254,
698
+ "median": 3.40593603515625,
699
+ "std": 0.0021956034435643246,
700
+ "min": 3.403377197265625,
701
+ "max": 3.409301025390625,
702
+ "p25": 3.4053486328125,
703
+ "p75": 3.408706298828125,
704
+ "p90": 3.409063134765625,
705
+ "p95": 3.409182080078125,
706
+ "p99": 3.409277236328125,
707
+ "unit": "seconds"
708
+ },
709
+ "time_to_first_token_seconds": {
710
+ "name": "time_to_first_token_seconds",
711
+ "measurements": [
712
+ 0.035305728912353514,
713
+ 0.034740577697753905,
714
+ 0.034734718322753905,
715
+ 0.034670974731445314,
716
+ 0.034718017578125
717
+ ],
718
+ "mean": 0.03483400344848633,
719
+ "median": 0.034734718322753905,
720
+ "std": 0.00023712524330228718,
721
+ "min": 0.034670974731445314,
722
+ "max": 0.035305728912353514,
723
+ "p25": 0.034718017578125,
724
+ "p75": 0.034740577697753905,
725
+ "p90": 0.03507966842651367,
726
+ "p95": 0.03519269866943359,
727
+ "p99": 0.03528312286376953,
728
+ "unit": "seconds"
729
+ },
730
+ "tokens_per_second": {
731
+ "name": "tokens_per_second",
732
+ "measurements": [
733
+ 29.3315255107291,
734
+ 29.360504415759653,
735
+ 29.382579186445447,
736
+ 29.336643064372804,
737
+ 29.365568927786796
738
+ ],
739
+ "mean": 29.35536422101876,
740
+ "median": 29.360504415759653,
741
+ "std": 0.018920770596402275,
742
+ "min": 29.3315255107291,
743
+ "max": 29.382579186445447,
744
+ "p25": 29.336643064372804,
745
+ "p75": 29.365568927786796,
746
+ "p90": 29.37577508298199,
747
+ "p95": 29.379177134713718,
748
+ "p99": 29.381898776099103,
749
+ "unit": "tokens/sec"
750
+ },
751
+ "time_per_output_token_seconds": {
752
+ "name": "time_per_output_token_seconds",
753
+ "measurements": [
754
+ 0.03409301025390625,
755
+ 0.0340593603515625,
756
+ 0.03403377197265625,
757
+ 0.03408706298828125,
758
+ 0.034053486328125
759
+ ],
760
+ "mean": 0.034065338378906254,
761
+ "median": 0.0340593603515625,
762
+ "std": 2.195603443564376e-05,
763
+ "min": 0.03403377197265625,
764
+ "max": 0.03409301025390625,
765
+ "p25": 0.034053486328125,
766
+ "p75": 0.03408706298828125,
767
+ "p90": 0.03409063134765625,
768
+ "p95": 0.03409182080078125,
769
+ "p99": 0.03409277236328125,
770
+ "unit": "seconds/token"
771
+ }
772
+ },
773
+ "gpu_metrics": {
774
+ "gpu_utilization_mean": 96.96511627906976,
775
+ "gpu_utilization_max": 98,
776
+ "gpu_utilization_min": 95,
777
+ "gpu_memory_used_mean": 13849,
778
+ "gpu_memory_used_max": 13849,
779
+ "gpu_memory_used_min": 13849,
780
+ "sample_count": 86,
781
+ "gpu_monitoring_status": "success"
782
+ }
783
+ },
784
+ {
785
+ "scenario_name": "compiled_compile_max-autotune_sdpa_default",
786
+ "metadata": {
787
+ "timestamp": "2025-10-05T16:47:42.652102",
788
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
789
+ "hardware_info": {
790
+ "gpu_name": "NVIDIA A10G",
791
+ "gpu_memory_total_mb": 23028,
792
+ "cpu_count": 16,
793
+ "memory_total_mb": 63607,
794
+ "python_version": "3.10.12",
795
+ "torch_version": "2.8.0+cu126",
796
+ "cuda_version": "12.6"
797
+ },
798
+ "config": {
799
+ "name": "compiled",
800
+ "model_id": "meta-llama/Llama-2-7b-hf",
801
+ "variant": "compiled",
802
+ "warmup_iterations": 3,
803
+ "measurement_iterations": 5,
804
+ "num_tokens_to_generate": 100,
805
+ "device": "cuda",
806
+ "torch_dtype": "float16",
807
+ "compile_mode": "max-autotune",
808
+ "compile_options": {},
809
+ "use_cache": true,
810
+ "batch_size": 1,
811
+ "sequence_length": null,
812
+ "attn_implementation": "sdpa",
813
+ "sdpa_backend": null,
814
+ "custom_params": {}
815
+ }
816
+ },
817
+ "measurements": {
818
+ "latency_seconds": {
819
+ "name": "latency_seconds",
820
+ "measurements": [
821
+ 3.685540771484375,
822
+ 3.69011865234375,
823
+ 3.68904296875,
824
+ 3.68791259765625,
825
+ 3.68812548828125
826
+ ],
827
+ "mean": 3.688148095703125,
828
+ "median": 3.68812548828125,
829
+ "std": 0.0015190787505077279,
830
+ "min": 3.685540771484375,
831
+ "max": 3.69011865234375,
832
+ "p25": 3.68791259765625,
833
+ "p75": 3.68904296875,
834
+ "p90": 3.6896883789062502,
835
+ "p95": 3.689903515625,
836
+ "p99": 3.690075625,
837
+ "unit": "seconds"
838
+ },
839
+ "time_to_first_token_seconds": {
840
+ "name": "time_to_first_token_seconds",
841
+ "measurements": [
842
+ 0.03792915344238281,
843
+ 0.037532608032226564,
844
+ 0.03800860977172851,
845
+ 0.03750201416015625,
846
+ 0.03744150543212891
847
+ ],
848
+ "mean": 0.037682778167724605,
849
+ "median": 0.037532608032226564,
850
+ "std": 0.00023677256559836882,
851
+ "min": 0.03744150543212891,
852
+ "max": 0.03800860977172851,
853
+ "p25": 0.03750201416015625,
854
+ "p75": 0.03792915344238281,
855
+ "p90": 0.037976827239990234,
856
+ "p95": 0.03799271850585937,
857
+ "p99": 0.038005431518554685,
858
+ "unit": "seconds"
859
+ },
860
+ "tokens_per_second": {
861
+ "name": "tokens_per_second",
862
+ "measurements": [
863
+ 27.13306030249785,
864
+ 27.099399618623586,
865
+ 27.107301499902054,
866
+ 27.11561007805668,
867
+ 27.114044876656912
868
+ ],
869
+ "mean": 27.113883275147412,
870
+ "median": 27.114044876656912,
871
+ "std": 0.01117016036604065,
872
+ "min": 27.099399618623586,
873
+ "max": 27.13306030249785,
874
+ "p25": 27.107301499902054,
875
+ "p75": 27.11561007805668,
876
+ "p90": 27.12608021272138,
877
+ "p95": 27.129570257609615,
878
+ "p99": 27.132362293520202,
879
+ "unit": "tokens/sec"
880
+ },
881
+ "time_per_output_token_seconds": {
882
+ "name": "time_per_output_token_seconds",
883
+ "measurements": [
884
+ 0.03685540771484375,
885
+ 0.036901186523437504,
886
+ 0.0368904296875,
887
+ 0.0368791259765625,
888
+ 0.0368812548828125
889
+ ],
890
+ "mean": 0.036881480957031254,
891
+ "median": 0.0368812548828125,
892
+ "std": 1.5190787505076791e-05,
893
+ "min": 0.03685540771484375,
894
+ "max": 0.036901186523437504,
895
+ "p25": 0.0368791259765625,
896
+ "p75": 0.0368904296875,
897
+ "p90": 0.0368968837890625,
898
+ "p95": 0.03689903515625,
899
+ "p99": 0.03690075625000001,
900
+ "unit": "seconds/token"
901
+ }
902
+ },
903
+ "gpu_metrics": {
904
+ "gpu_utilization_mean": 90.68817204301075,
905
+ "gpu_utilization_max": 93,
906
+ "gpu_utilization_min": 88,
907
+ "gpu_memory_used_mean": 13845,
908
+ "gpu_memory_used_max": 13845,
909
+ "gpu_memory_used_min": 13845,
910
+ "sample_count": 93,
911
+ "gpu_monitoring_status": "success"
912
+ }
913
+ },
914
+ {
915
+ "scenario_name": "compiled_compile_max-autotune_sdpa_math",
916
+ "metadata": {
917
+ "timestamp": "2025-10-05T16:48:17.314600",
918
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
919
+ "hardware_info": {
920
+ "gpu_name": "NVIDIA A10G",
921
+ "gpu_memory_total_mb": 23028,
922
+ "cpu_count": 16,
923
+ "memory_total_mb": 63607,
924
+ "python_version": "3.10.12",
925
+ "torch_version": "2.8.0+cu126",
926
+ "cuda_version": "12.6"
927
+ },
928
+ "config": {
929
+ "name": "compiled",
930
+ "model_id": "meta-llama/Llama-2-7b-hf",
931
+ "variant": "compiled",
932
+ "warmup_iterations": 3,
933
+ "measurement_iterations": 5,
934
+ "num_tokens_to_generate": 100,
935
+ "device": "cuda",
936
+ "torch_dtype": "float16",
937
+ "compile_mode": "max-autotune",
938
+ "compile_options": {},
939
+ "use_cache": true,
940
+ "batch_size": 1,
941
+ "sequence_length": null,
942
+ "attn_implementation": "sdpa",
943
+ "sdpa_backend": "math",
944
+ "custom_params": {}
945
+ }
946
+ },
947
+ "measurements": {
948
+ "latency_seconds": {
949
+ "name": "latency_seconds",
950
+ "measurements": [
951
+ 3.86907470703125,
952
+ 3.87136767578125,
953
+ 3.875137451171875,
954
+ 3.90616748046875,
955
+ 3.85316552734375
956
+ ],
957
+ "mean": 3.8749825683593757,
958
+ "median": 3.87136767578125,
959
+ "std": 0.01730015193260598,
960
+ "min": 3.85316552734375,
961
+ "max": 3.90616748046875,
962
+ "p25": 3.86907470703125,
963
+ "p75": 3.875137451171875,
964
+ "p90": 3.8937554687500002,
965
+ "p95": 3.899961474609375,
966
+ "p99": 3.904926279296875,
967
+ "unit": "seconds"
968
+ },
969
+ "time_to_first_token_seconds": {
970
+ "name": "time_to_first_token_seconds",
971
+ "measurements": [
972
+ 0.04121488189697266,
973
+ 0.04030598449707031,
974
+ 0.040963680267333984,
975
+ 0.041119617462158205,
976
+ 0.03978636932373047
977
+ ],
978
+ "mean": 0.040678106689453125,
979
+ "median": 0.040963680267333984,
980
+ "std": 0.0005474088252295032,
981
+ "min": 0.03978636932373047,
982
+ "max": 0.04121488189697266,
983
+ "p25": 0.04030598449707031,
984
+ "p75": 0.041119617462158205,
985
+ "p90": 0.041176776123046877,
986
+ "p95": 0.04119582901000977,
987
+ "p99": 0.04121107131958008,
988
+ "unit": "seconds"
989
+ },
990
+ "tokens_per_second": {
991
+ "name": "tokens_per_second",
992
+ "measurements": [
993
+ 25.845972893278724,
994
+ 25.830664606099393,
995
+ 25.80553625775497,
996
+ 25.60054081142464,
997
+ 25.952687287985995
998
+ ],
999
+ "mean": 25.807080371308746,
1000
+ "median": 25.830664606099393,
1001
+ "std": 0.11483134821457781,
1002
+ "min": 25.60054081142464,
1003
+ "max": 25.952687287985995,
1004
+ "p25": 25.80553625775497,
1005
+ "p75": 25.845972893278724,
1006
+ "p90": 25.910001530103088,
1007
+ "p95": 25.93134440904454,
1008
+ "p99": 25.948418712197704,
1009
+ "unit": "tokens/sec"
1010
+ },
1011
+ "time_per_output_token_seconds": {
1012
+ "name": "time_per_output_token_seconds",
1013
+ "measurements": [
1014
+ 0.038690747070312496,
1015
+ 0.038713676757812496,
1016
+ 0.03875137451171875,
1017
+ 0.0390616748046875,
1018
+ 0.0385316552734375
1019
+ ],
1020
+ "mean": 0.038749825683593746,
1021
+ "median": 0.038713676757812496,
1022
+ "std": 0.00017300151932606096,
1023
+ "min": 0.0385316552734375,
1024
+ "max": 0.0390616748046875,
1025
+ "p25": 0.038690747070312496,
1026
+ "p75": 0.03875137451171875,
1027
+ "p90": 0.0389375546875,
1028
+ "p95": 0.03899961474609375,
1029
+ "p99": 0.039049262792968754,
1030
+ "unit": "seconds/token"
1031
+ }
1032
+ },
1033
+ "gpu_metrics": {
1034
+ "gpu_utilization_mean": 89.39795918367346,
1035
+ "gpu_utilization_max": 93,
1036
+ "gpu_utilization_min": 82,
1037
+ "gpu_memory_used_mean": 13845,
1038
+ "gpu_memory_used_max": 13845,
1039
+ "gpu_memory_used_min": 13845,
1040
+ "sample_count": 98,
1041
+ "gpu_monitoring_status": "success"
1042
+ }
1043
+ },
1044
+ {
1045
+ "scenario_name": "compiled_compile_max-autotune_sdpa_efficient_attention",
1046
+ "metadata": {
1047
+ "timestamp": "2025-10-05T16:48:58.291199",
1048
+ "commit_id": "e11a00a16f925b7d3b52f5007bdce3464edb361f",
1049
+ "hardware_info": {
1050
+ "gpu_name": "NVIDIA A10G",
1051
+ "gpu_memory_total_mb": 23028,
1052
+ "cpu_count": 16,
1053
+ "memory_total_mb": 63607,
1054
+ "python_version": "3.10.12",
1055
+ "torch_version": "2.8.0+cu126",
1056
+ "cuda_version": "12.6"
1057
+ },
1058
+ "config": {
1059
+ "name": "compiled",
1060
+ "model_id": "meta-llama/Llama-2-7b-hf",
1061
+ "variant": "compiled",
1062
+ "warmup_iterations": 3,
1063
+ "measurement_iterations": 5,
1064
+ "num_tokens_to_generate": 100,
1065
+ "device": "cuda",
1066
+ "torch_dtype": "float16",
1067
+ "compile_mode": "max-autotune",
1068
+ "compile_options": {},
1069
+ "use_cache": true,
1070
+ "batch_size": 1,
1071
+ "sequence_length": null,
1072
+ "attn_implementation": "sdpa",
1073
+ "sdpa_backend": "efficient_attention",
1074
+ "custom_params": {}
1075
+ }
1076
+ },
1077
+ "measurements": {
1078
+ "latency_seconds": {
1079
+ "name": "latency_seconds",
1080
+ "measurements": [
1081
+ 3.69019482421875,
1082
+ 3.682122802734375,
1083
+ 3.6891953125,
1084
+ 3.690829833984375,
1085
+ 3.69011474609375
1086
+ ],
1087
+ "mean": 3.688491503906249,
1088
+ "median": 3.69011474609375,
1089
+ "std": 0.003226753910473955,
1090
+ "min": 3.682122802734375,
1091
+ "max": 3.690829833984375,
1092
+ "p25": 3.6891953125,
1093
+ "p75": 3.69019482421875,
1094
+ "p90": 3.690575830078125,
1095
+ "p95": 3.69070283203125,
1096
+ "p99": 3.6908044335937498,
1097
+ "unit": "seconds"
1098
+ },
1099
+ "time_to_first_token_seconds": {
1100
+ "name": "time_to_first_token_seconds",
1101
+ "measurements": [
1102
+ 0.038126495361328124,
1103
+ 0.03752556610107422,
1104
+ 0.037846046447753905,
1105
+ 0.037650623321533204,
1106
+ 0.03778144073486328
1107
+ ],
1108
+ "mean": 0.03778603439331055,
1109
+ "median": 0.03778144073486328,
1110
+ "std": 0.00020283791989680056,
1111
+ "min": 0.03752556610107422,
1112
+ "max": 0.038126495361328124,
1113
+ "p25": 0.037650623321533204,
1114
+ "p75": 0.037846046447753905,
1115
+ "p90": 0.03801431579589844,
1116
+ "p95": 0.03807040557861328,
1117
+ "p99": 0.03811527740478515,
1118
+ "unit": "seconds"
1119
+ },
1120
+ "tokens_per_second": {
1121
+ "name": "tokens_per_second",
1122
+ "measurements": [
1123
+ 27.0988402410897,
1124
+ 27.15824684764429,
1125
+ 27.10618211542575,
1126
+ 27.09417786732439,
1127
+ 27.099428305273474
1128
+ ],
1129
+ "mean": 27.11137507535152,
1130
+ "median": 27.099428305273474,
1131
+ "std": 0.023746668060396277,
1132
+ "min": 27.09417786732439,
1133
+ "max": 27.15824684764429,
1134
+ "p25": 27.0988402410897,
1135
+ "p75": 27.10618211542575,
1136
+ "p90": 27.137420954756873,
1137
+ "p95": 27.14783390120058,
1138
+ "p99": 27.15616425835555,
1139
+ "unit": "tokens/sec"
1140
+ },
1141
+ "time_per_output_token_seconds": {
1142
+ "name": "time_per_output_token_seconds",
1143
+ "measurements": [
1144
+ 0.0369019482421875,
1145
+ 0.03682122802734375,
1146
+ 0.036891953125,
1147
+ 0.03690829833984375,
1148
+ 0.0369011474609375
1149
+ ],
1150
+ "mean": 0.03688491503906251,
1151
+ "median": 0.0369011474609375,
1152
+ "std": 3.22675391047406e-05,
1153
+ "min": 0.03682122802734375,
1154
+ "max": 0.03690829833984375,
1155
+ "p25": 0.036891953125,
1156
+ "p75": 0.0369019482421875,
1157
+ "p90": 0.03690575830078125,
1158
+ "p95": 0.0369070283203125,
1159
+ "p99": 0.0369080443359375,
1160
+ "unit": "seconds/token"
1161
+ }
1162
+ },
1163
+ "gpu_metrics": {
1164
+ "gpu_utilization_mean": 90.65591397849462,
1165
+ "gpu_utilization_max": 94,
1166
+ "gpu_utilization_min": 88,
1167
+ "gpu_memory_used_mean": 13845,
1168
+ "gpu_memory_used_max": 13845,
1169
+ "gpu_memory_used_min": 13845,
1170
+ "sample_count": 93,
1171
+ "gpu_monitoring_status": "success"
1172
+ }
1173
+ }
1174
+ ]
1175
+ }