hf-transformers-bot commited on
Commit
cd3bc23
·
verified ·
1 Parent(s): ba350f3

Upload benchmark results for run 18382742695

Browse files
2025-10-09/18382742695/benchmark_results/Llama-2-7b-hf/Llama-2-7b-hf_benchmark_20251009_165612.json ADDED
@@ -0,0 +1,1121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_name": "Llama-2-7b-hf",
3
+ "benchmark_scenarios": [
4
+ {
5
+ "scenario_name": "eager_eager_attn",
6
+ "metadata": {
7
+ "timestamp": "2025-10-09T16:49:24.488978",
8
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
9
+ "hardware_info": {
10
+ "gpu_name": "unknown",
11
+ "gpu_memory_total_mb": 0,
12
+ "cpu_count": 160,
13
+ "memory_total_mb": 1290069,
14
+ "python_version": "3.12.10",
15
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
16
+ "cuda_version": null
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
+ 1.7057410888671876,
42
+ 1.6808692626953126,
43
+ 1.860894775390625,
44
+ 1.688207275390625,
45
+ 1.720949462890625
46
+ ],
47
+ "mean": 1.7313323730468753,
48
+ "median": 1.7057410888671876,
49
+ "std": 0.06626527643430928,
50
+ "min": 1.6808692626953126,
51
+ "max": 1.860894775390625,
52
+ "p25": 1.688207275390625,
53
+ "p75": 1.720949462890625,
54
+ "p90": 1.804916650390625,
55
+ "p95": 1.832905712890625,
56
+ "p99": 1.855296962890625,
57
+ "unit": "seconds"
58
+ },
59
+ "time_to_first_token_seconds": {
60
+ "name": "time_to_first_token_seconds",
61
+ "measurements": [
62
+ 0.01838970756530762,
63
+ 0.014546339035034179,
64
+ 0.020507736206054686,
65
+ 0.02193131637573242,
66
+ 0.027124168395996094
67
+ ],
68
+ "mean": 0.020499853515625,
69
+ "median": 0.020507736206054686,
70
+ "std": 0.004143132783129854,
71
+ "min": 0.014546339035034179,
72
+ "max": 0.027124168395996094,
73
+ "p25": 0.01838970756530762,
74
+ "p75": 0.02193131637573242,
75
+ "p90": 0.025047027587890625,
76
+ "p95": 0.02608559799194336,
77
+ "p99": 0.026916454315185547,
78
+ "unit": "seconds"
79
+ },
80
+ "tokens_per_second": {
81
+ "name": "tokens_per_second",
82
+ "measurements": [
83
+ 58.62554443500669,
84
+ 59.49302674477353,
85
+ 53.737589745776326,
86
+ 59.234432559154534,
87
+ 58.10745879314383
88
+ ],
89
+ "mean": 57.839610455570984,
90
+ "median": 58.62554443500669,
91
+ "std": 2.106908777408363,
92
+ "min": 53.737589745776326,
93
+ "max": 59.49302674477353,
94
+ "p25": 58.10745879314383,
95
+ "p75": 59.234432559154534,
96
+ "p90": 59.38958907052593,
97
+ "p95": 59.44130790764973,
98
+ "p99": 59.48268297734877,
99
+ "unit": "tokens/sec"
100
+ },
101
+ "time_per_output_token_seconds": {
102
+ "name": "time_per_output_token_seconds",
103
+ "measurements": [
104
+ 0.017057410888671874,
105
+ 0.016808692626953125,
106
+ 0.01860894775390625,
107
+ 0.01688207275390625,
108
+ 0.01720949462890625
109
+ ],
110
+ "mean": 0.01731332373046875,
111
+ "median": 0.017057410888671874,
112
+ "std": 0.0006626527643430928,
113
+ "min": 0.016808692626953125,
114
+ "max": 0.01860894775390625,
115
+ "p25": 0.01688207275390625,
116
+ "p75": 0.01720949462890625,
117
+ "p90": 0.01804916650390625,
118
+ "p95": 0.01832905712890625,
119
+ "p99": 0.01855296962890625,
120
+ "unit": "seconds/token"
121
+ }
122
+ },
123
+ "gpu_metrics": {
124
+ "gpu_monitoring_status": "disabled",
125
+ "gpu_monitoring_reason": "no_gpus_available"
126
+ }
127
+ },
128
+ {
129
+ "scenario_name": "compiled_compile_max-autotune_eager_attn",
130
+ "metadata": {
131
+ "timestamp": "2025-10-09T16:49:59.932130",
132
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
133
+ "hardware_info": {
134
+ "gpu_name": "unknown",
135
+ "gpu_memory_total_mb": 0,
136
+ "cpu_count": 160,
137
+ "memory_total_mb": 1290069,
138
+ "python_version": "3.12.10",
139
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
140
+ "cuda_version": null
141
+ },
142
+ "config": {
143
+ "name": "compiled",
144
+ "model_id": "meta-llama/Llama-2-7b-hf",
145
+ "variant": "compiled",
146
+ "warmup_iterations": 3,
147
+ "measurement_iterations": 5,
148
+ "num_tokens_to_generate": 100,
149
+ "device": "cuda",
150
+ "torch_dtype": "float16",
151
+ "compile_mode": "max-autotune",
152
+ "compile_options": {},
153
+ "use_cache": true,
154
+ "batch_size": 1,
155
+ "sequence_length": null,
156
+ "attn_implementation": "eager",
157
+ "sdpa_backend": null,
158
+ "custom_params": {}
159
+ }
160
+ },
161
+ "measurements": {
162
+ "latency_seconds": {
163
+ "name": "latency_seconds",
164
+ "measurements": [
165
+ 3.081891357421875,
166
+ 3.066642822265625,
167
+ 3.05245263671875,
168
+ 3.21873291015625,
169
+ 3.291984619140625
170
+ ],
171
+ "mean": 3.1423408691406247,
172
+ "median": 3.081891357421875,
173
+ "std": 0.0955962634628043,
174
+ "min": 3.05245263671875,
175
+ "max": 3.291984619140625,
176
+ "p25": 3.066642822265625,
177
+ "p75": 3.21873291015625,
178
+ "p90": 3.262683935546875,
179
+ "p95": 3.27733427734375,
180
+ "p99": 3.2890545507812496,
181
+ "unit": "seconds"
182
+ },
183
+ "time_to_first_token_seconds": {
184
+ "name": "time_to_first_token_seconds",
185
+ "measurements": [
186
+ 0.017425079345703125,
187
+ 0.0181048641204834,
188
+ 0.01802862548828125,
189
+ 0.023720348358154297,
190
+ 0.018767732620239258
191
+ ],
192
+ "mean": 0.019209329986572266,
193
+ "median": 0.0181048641204834,
194
+ "std": 0.002295288796373339,
195
+ "min": 0.017425079345703125,
196
+ "max": 0.023720348358154297,
197
+ "p25": 0.01802862548828125,
198
+ "p75": 0.018767732620239258,
199
+ "p90": 0.021739302062988283,
200
+ "p95": 0.022729825210571288,
201
+ "p99": 0.023522243728637694,
202
+ "unit": "seconds"
203
+ },
204
+ "tokens_per_second": {
205
+ "name": "tokens_per_second",
206
+ "measurements": [
207
+ 32.44760713552667,
208
+ 32.60894919810725,
209
+ 32.7605410800069,
210
+ 31.068126120208465,
211
+ 30.376812643221
212
+ ],
213
+ "mean": 31.852407235414056,
214
+ "median": 32.44760713552667,
215
+ "std": 0.9532892541387661,
216
+ "min": 30.376812643221,
217
+ "max": 32.7605410800069,
218
+ "p25": 31.068126120208465,
219
+ "p75": 32.60894919810725,
220
+ "p90": 32.69990432724704,
221
+ "p95": 32.73022270362697,
222
+ "p99": 32.75447740473091,
223
+ "unit": "tokens/sec"
224
+ },
225
+ "time_per_output_token_seconds": {
226
+ "name": "time_per_output_token_seconds",
227
+ "measurements": [
228
+ 0.03081891357421875,
229
+ 0.030666428222656247,
230
+ 0.0305245263671875,
231
+ 0.0321873291015625,
232
+ 0.03291984619140625
233
+ ],
234
+ "mean": 0.03142340869140625,
235
+ "median": 0.03081891357421875,
236
+ "std": 0.0009559626346280432,
237
+ "min": 0.0305245263671875,
238
+ "max": 0.03291984619140625,
239
+ "p25": 0.030666428222656247,
240
+ "p75": 0.0321873291015625,
241
+ "p90": 0.03262683935546875,
242
+ "p95": 0.0327733427734375,
243
+ "p99": 0.0328905455078125,
244
+ "unit": "seconds/token"
245
+ }
246
+ },
247
+ "gpu_metrics": {
248
+ "gpu_monitoring_status": "disabled",
249
+ "gpu_monitoring_reason": "no_gpus_available"
250
+ }
251
+ },
252
+ {
253
+ "scenario_name": "eager_sdpa_default",
254
+ "metadata": {
255
+ "timestamp": "2025-10-09T16:51:03.953386",
256
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
257
+ "hardware_info": {
258
+ "gpu_name": "unknown",
259
+ "gpu_memory_total_mb": 0,
260
+ "cpu_count": 160,
261
+ "memory_total_mb": 1290069,
262
+ "python_version": "3.12.10",
263
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
264
+ "cuda_version": null
265
+ },
266
+ "config": {
267
+ "name": "eager",
268
+ "model_id": "meta-llama/Llama-2-7b-hf",
269
+ "variant": "eager",
270
+ "warmup_iterations": 3,
271
+ "measurement_iterations": 5,
272
+ "num_tokens_to_generate": 100,
273
+ "device": "cuda",
274
+ "torch_dtype": "float16",
275
+ "compile_mode": null,
276
+ "compile_options": {},
277
+ "use_cache": true,
278
+ "batch_size": 1,
279
+ "sequence_length": null,
280
+ "attn_implementation": "sdpa",
281
+ "sdpa_backend": null,
282
+ "custom_params": {}
283
+ }
284
+ },
285
+ "measurements": {
286
+ "latency_seconds": {
287
+ "name": "latency_seconds",
288
+ "measurements": [
289
+ 1.23875634765625,
290
+ 1.2362845458984375,
291
+ 1.2198138427734375,
292
+ 1.3482608642578124,
293
+ 1.4902548828125
294
+ ],
295
+ "mean": 1.3066740966796875,
296
+ "median": 1.23875634765625,
297
+ "std": 0.10251202609937377,
298
+ "min": 1.2198138427734375,
299
+ "max": 1.4902548828125,
300
+ "p25": 1.2362845458984375,
301
+ "p75": 1.3482608642578124,
302
+ "p90": 1.433457275390625,
303
+ "p95": 1.4618560791015625,
304
+ "p99": 1.4845751220703125,
305
+ "unit": "seconds"
306
+ },
307
+ "time_to_first_token_seconds": {
308
+ "name": "time_to_first_token_seconds",
309
+ "measurements": [
310
+ 0.013268726348876953,
311
+ 0.013707356452941894,
312
+ 0.013279444694519043,
313
+ 0.013177566528320313,
314
+ 0.016668214797973632
315
+ ],
316
+ "mean": 0.014020261764526365,
317
+ "median": 0.013279444694519043,
318
+ "std": 0.0013366611892827816,
319
+ "min": 0.013177566528320313,
320
+ "max": 0.016668214797973632,
321
+ "p25": 0.013268726348876953,
322
+ "p75": 0.013707356452941894,
323
+ "p90": 0.015483871459960937,
324
+ "p95": 0.016076043128967283,
325
+ "p99": 0.016549780464172364,
326
+ "unit": "seconds"
327
+ },
328
+ "tokens_per_second": {
329
+ "name": "tokens_per_second",
330
+ "measurements": [
331
+ 80.72612518934967,
332
+ 80.8875273348399,
333
+ 81.97972222764284,
334
+ 74.16962299432149,
335
+ 67.10261523268684
336
+ ],
337
+ "mean": 76.97312259576816,
338
+ "median": 80.72612518934967,
339
+ "std": 5.652612819264003,
340
+ "min": 67.10261523268684,
341
+ "max": 81.97972222764284,
342
+ "p25": 74.16962299432149,
343
+ "p75": 80.8875273348399,
344
+ "p90": 81.54284427052167,
345
+ "p95": 81.76128324908225,
346
+ "p99": 81.93603443193072,
347
+ "unit": "tokens/sec"
348
+ },
349
+ "time_per_output_token_seconds": {
350
+ "name": "time_per_output_token_seconds",
351
+ "measurements": [
352
+ 0.0123875634765625,
353
+ 0.012362845458984376,
354
+ 0.012198138427734375,
355
+ 0.013482608642578124,
356
+ 0.014902548828124999
357
+ ],
358
+ "mean": 0.013066740966796875,
359
+ "median": 0.0123875634765625,
360
+ "std": 0.0010251202609937373,
361
+ "min": 0.012198138427734375,
362
+ "max": 0.014902548828124999,
363
+ "p25": 0.012362845458984376,
364
+ "p75": 0.013482608642578124,
365
+ "p90": 0.01433457275390625,
366
+ "p95": 0.014618560791015624,
367
+ "p99": 0.014845751220703123,
368
+ "unit": "seconds/token"
369
+ }
370
+ },
371
+ "gpu_metrics": {
372
+ "gpu_monitoring_status": "disabled",
373
+ "gpu_monitoring_reason": "no_gpus_available"
374
+ }
375
+ },
376
+ {
377
+ "scenario_name": "eager_sdpa_math",
378
+ "metadata": {
379
+ "timestamp": "2025-10-09T16:51:38.828527",
380
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
381
+ "hardware_info": {
382
+ "gpu_name": "unknown",
383
+ "gpu_memory_total_mb": 0,
384
+ "cpu_count": 160,
385
+ "memory_total_mb": 1290069,
386
+ "python_version": "3.12.10",
387
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
388
+ "cuda_version": null
389
+ },
390
+ "config": {
391
+ "name": "eager",
392
+ "model_id": "meta-llama/Llama-2-7b-hf",
393
+ "variant": "eager",
394
+ "warmup_iterations": 3,
395
+ "measurement_iterations": 5,
396
+ "num_tokens_to_generate": 100,
397
+ "device": "cuda",
398
+ "torch_dtype": "float16",
399
+ "compile_mode": null,
400
+ "compile_options": {},
401
+ "use_cache": true,
402
+ "batch_size": 1,
403
+ "sequence_length": null,
404
+ "attn_implementation": "sdpa",
405
+ "sdpa_backend": "math",
406
+ "custom_params": {}
407
+ }
408
+ },
409
+ "measurements": {
410
+ "latency_seconds": {
411
+ "name": "latency_seconds",
412
+ "measurements": [
413
+ 1.6838673095703125,
414
+ 1.990463623046875,
415
+ 1.669430419921875,
416
+ 1.99741650390625,
417
+ 2.0392139892578127
418
+ ],
419
+ "mean": 1.8760783691406253,
420
+ "median": 1.990463623046875,
421
+ "std": 0.16374897016352494,
422
+ "min": 1.669430419921875,
423
+ "max": 2.0392139892578127,
424
+ "p25": 1.6838673095703125,
425
+ "p75": 1.99741650390625,
426
+ "p90": 2.0224949951171878,
427
+ "p95": 2.0308544921875002,
428
+ "p99": 2.0375420898437504,
429
+ "unit": "seconds"
430
+ },
431
+ "time_to_first_token_seconds": {
432
+ "name": "time_to_first_token_seconds",
433
+ "measurements": [
434
+ 0.021613391876220703,
435
+ 0.023089960098266603,
436
+ 0.01894112777709961,
437
+ 0.021053644180297853,
438
+ 0.021367597579956054
439
+ ],
440
+ "mean": 0.021213144302368165,
441
+ "median": 0.021367597579956054,
442
+ "std": 0.0013337159253079508,
443
+ "min": 0.01894112777709961,
444
+ "max": 0.023089960098266603,
445
+ "p25": 0.021053644180297853,
446
+ "p75": 0.021613391876220703,
447
+ "p90": 0.02249933280944824,
448
+ "p95": 0.022794646453857422,
449
+ "p99": 0.023030897369384767,
450
+ "unit": "seconds"
451
+ },
452
+ "tokens_per_second": {
453
+ "name": "tokens_per_second",
454
+ "measurements": [
455
+ 59.38710219721404,
456
+ 50.23955165125116,
457
+ 59.90066959764621,
458
+ 50.06467094090536,
459
+ 49.038502347855975
460
+ ],
461
+ "mean": 53.72609934697455,
462
+ "median": 50.23955165125116,
463
+ "std": 4.8519668070663675,
464
+ "min": 49.038502347855975,
465
+ "max": 59.90066959764621,
466
+ "p25": 50.06467094090536,
467
+ "p75": 59.38710219721404,
468
+ "p90": 59.69524263747334,
469
+ "p95": 59.79795611755977,
470
+ "p99": 59.88012690162892,
471
+ "unit": "tokens/sec"
472
+ },
473
+ "time_per_output_token_seconds": {
474
+ "name": "time_per_output_token_seconds",
475
+ "measurements": [
476
+ 0.016838673095703124,
477
+ 0.01990463623046875,
478
+ 0.01669430419921875,
479
+ 0.0199741650390625,
480
+ 0.020392139892578126
481
+ ],
482
+ "mean": 0.01876078369140625,
483
+ "median": 0.01990463623046875,
484
+ "std": 0.0016374897016352497,
485
+ "min": 0.01669430419921875,
486
+ "max": 0.020392139892578126,
487
+ "p25": 0.016838673095703124,
488
+ "p75": 0.0199741650390625,
489
+ "p90": 0.020224949951171876,
490
+ "p95": 0.020308544921875,
491
+ "p99": 0.0203754208984375,
492
+ "unit": "seconds/token"
493
+ }
494
+ },
495
+ "gpu_metrics": {
496
+ "gpu_monitoring_status": "disabled",
497
+ "gpu_monitoring_reason": "no_gpus_available"
498
+ }
499
+ },
500
+ {
501
+ "scenario_name": "eager_sdpa_flash_attention",
502
+ "metadata": {
503
+ "timestamp": "2025-10-09T16:52:32.329744",
504
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
505
+ "hardware_info": {
506
+ "gpu_name": "unknown",
507
+ "gpu_memory_total_mb": 0,
508
+ "cpu_count": 160,
509
+ "memory_total_mb": 1290069,
510
+ "python_version": "3.12.10",
511
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
512
+ "cuda_version": null
513
+ },
514
+ "config": {
515
+ "name": "eager",
516
+ "model_id": "meta-llama/Llama-2-7b-hf",
517
+ "variant": "eager",
518
+ "warmup_iterations": 3,
519
+ "measurement_iterations": 5,
520
+ "num_tokens_to_generate": 100,
521
+ "device": "cuda",
522
+ "torch_dtype": "float16",
523
+ "compile_mode": null,
524
+ "compile_options": {},
525
+ "use_cache": true,
526
+ "batch_size": 1,
527
+ "sequence_length": null,
528
+ "attn_implementation": "sdpa",
529
+ "sdpa_backend": "flash_attention",
530
+ "custom_params": {}
531
+ }
532
+ },
533
+ "measurements": {
534
+ "latency_seconds": {
535
+ "name": "latency_seconds",
536
+ "measurements": [
537
+ 1.7609259033203124,
538
+ 1.5393101806640626,
539
+ 1.338537841796875,
540
+ 1.2446170654296875,
541
+ 1.32676416015625
542
+ ],
543
+ "mean": 1.4420310302734374,
544
+ "median": 1.338537841796875,
545
+ "std": 0.1866155074367034,
546
+ "min": 1.2446170654296875,
547
+ "max": 1.7609259033203124,
548
+ "p25": 1.32676416015625,
549
+ "p75": 1.5393101806640626,
550
+ "p90": 1.6722796142578125,
551
+ "p95": 1.7166027587890624,
552
+ "p99": 1.7520612744140625,
553
+ "unit": "seconds"
554
+ },
555
+ "time_to_first_token_seconds": {
556
+ "name": "time_to_first_token_seconds",
557
+ "measurements": [
558
+ 0.015766193389892578,
559
+ 0.020846532821655273,
560
+ 0.017071447372436523,
561
+ 0.01365507698059082,
562
+ 0.015696075439453124
563
+ ],
564
+ "mean": 0.016607065200805664,
565
+ "median": 0.015766193389892578,
566
+ "std": 0.002384950259838785,
567
+ "min": 0.01365507698059082,
568
+ "max": 0.020846532821655273,
569
+ "p25": 0.015696075439453124,
570
+ "p75": 0.017071447372436523,
571
+ "p90": 0.01933649864196777,
572
+ "p95": 0.020091515731811522,
573
+ "p99": 0.020695529403686522,
574
+ "unit": "seconds"
575
+ },
576
+ "tokens_per_second": {
577
+ "name": "tokens_per_second",
578
+ "measurements": [
579
+ 56.78830654455425,
580
+ 64.96416463435571,
581
+ 74.70838468470818,
582
+ 80.34599779931213,
583
+ 75.37134556621066
584
+ ],
585
+ "mean": 70.43563984582819,
586
+ "median": 74.70838468470818,
587
+ "std": 8.450082619813758,
588
+ "min": 56.78830654455425,
589
+ "max": 80.34599779931213,
590
+ "p25": 64.96416463435571,
591
+ "p75": 75.37134556621066,
592
+ "p90": 78.35613690607154,
593
+ "p95": 79.35106735269184,
594
+ "p99": 80.14701170998808,
595
+ "unit": "tokens/sec"
596
+ },
597
+ "time_per_output_token_seconds": {
598
+ "name": "time_per_output_token_seconds",
599
+ "measurements": [
600
+ 0.017609259033203123,
601
+ 0.015393101806640625,
602
+ 0.01338537841796875,
603
+ 0.012446170654296875,
604
+ 0.013267641601562501
605
+ ],
606
+ "mean": 0.014420310302734374,
607
+ "median": 0.01338537841796875,
608
+ "std": 0.0018661550743670336,
609
+ "min": 0.012446170654296875,
610
+ "max": 0.017609259033203123,
611
+ "p25": 0.013267641601562501,
612
+ "p75": 0.015393101806640625,
613
+ "p90": 0.016722796142578124,
614
+ "p95": 0.017166027587890622,
615
+ "p99": 0.017520612744140622,
616
+ "unit": "seconds/token"
617
+ }
618
+ },
619
+ "gpu_metrics": {
620
+ "gpu_monitoring_status": "disabled",
621
+ "gpu_monitoring_reason": "no_gpus_available"
622
+ }
623
+ },
624
+ {
625
+ "scenario_name": "eager_sdpa_efficient_attention",
626
+ "metadata": {
627
+ "timestamp": "2025-10-09T16:53:19.684671",
628
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
629
+ "hardware_info": {
630
+ "gpu_name": "unknown",
631
+ "gpu_memory_total_mb": 0,
632
+ "cpu_count": 160,
633
+ "memory_total_mb": 1290069,
634
+ "python_version": "3.12.10",
635
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
636
+ "cuda_version": null
637
+ },
638
+ "config": {
639
+ "name": "eager",
640
+ "model_id": "meta-llama/Llama-2-7b-hf",
641
+ "variant": "eager",
642
+ "warmup_iterations": 3,
643
+ "measurement_iterations": 5,
644
+ "num_tokens_to_generate": 100,
645
+ "device": "cuda",
646
+ "torch_dtype": "float16",
647
+ "compile_mode": null,
648
+ "compile_options": {},
649
+ "use_cache": true,
650
+ "batch_size": 1,
651
+ "sequence_length": null,
652
+ "attn_implementation": "sdpa",
653
+ "sdpa_backend": "efficient_attention",
654
+ "custom_params": {}
655
+ }
656
+ },
657
+ "measurements": {
658
+ "latency_seconds": {
659
+ "name": "latency_seconds",
660
+ "measurements": [
661
+ 1.2402978515625,
662
+ 1.4074083251953124,
663
+ 1.2171650390625,
664
+ 1.48508251953125,
665
+ 1.1188226318359376
666
+ ],
667
+ "mean": 1.2937552734375,
668
+ "median": 1.2402978515625,
669
+ "std": 0.1333029034376388,
670
+ "min": 1.1188226318359376,
671
+ "max": 1.48508251953125,
672
+ "p25": 1.2171650390625,
673
+ "p75": 1.4074083251953124,
674
+ "p90": 1.454012841796875,
675
+ "p95": 1.4695476806640626,
676
+ "p99": 1.4819755517578126,
677
+ "unit": "seconds"
678
+ },
679
+ "time_to_first_token_seconds": {
680
+ "name": "time_to_first_token_seconds",
681
+ "measurements": [
682
+ 0.014814852714538574,
683
+ 0.016888652801513672,
684
+ 0.012810733795166016,
685
+ 0.01708628845214844,
686
+ 0.0323662109375
687
+ ],
688
+ "mean": 0.01879334774017334,
689
+ "median": 0.016888652801513672,
690
+ "std": 0.0069625372082241545,
691
+ "min": 0.012810733795166016,
692
+ "max": 0.0323662109375,
693
+ "p25": 0.014814852714538574,
694
+ "p75": 0.01708628845214844,
695
+ "p90": 0.026254241943359378,
696
+ "p95": 0.029310226440429684,
697
+ "p99": 0.031755014038085935,
698
+ "unit": "seconds"
699
+ },
700
+ "tokens_per_second": {
701
+ "name": "tokens_per_second",
702
+ "measurements": [
703
+ 80.62579474278876,
704
+ 71.05258524467129,
705
+ 82.15812711563196,
706
+ 67.33632554746109,
707
+ 89.37967212542394
708
+ ],
709
+ "mean": 78.11050095519542,
710
+ "median": 80.62579474278876,
711
+ "std": 7.9448250972561505,
712
+ "min": 67.33632554746109,
713
+ "max": 89.37967212542394,
714
+ "p25": 71.05258524467129,
715
+ "p75": 82.15812711563196,
716
+ "p90": 86.49105412150715,
717
+ "p95": 87.93536312346554,
718
+ "p99": 89.09081032503225,
719
+ "unit": "tokens/sec"
720
+ },
721
+ "time_per_output_token_seconds": {
722
+ "name": "time_per_output_token_seconds",
723
+ "measurements": [
724
+ 0.012402978515625001,
725
+ 0.014074083251953124,
726
+ 0.012171650390625,
727
+ 0.0148508251953125,
728
+ 0.011188226318359376
729
+ ],
730
+ "mean": 0.012937552734375002,
731
+ "median": 0.012402978515625001,
732
+ "std": 0.0013330290343763875,
733
+ "min": 0.011188226318359376,
734
+ "max": 0.0148508251953125,
735
+ "p25": 0.012171650390625,
736
+ "p75": 0.014074083251953124,
737
+ "p90": 0.01454012841796875,
738
+ "p95": 0.014695476806640625,
739
+ "p99": 0.014819755517578125,
740
+ "unit": "seconds/token"
741
+ }
742
+ },
743
+ "gpu_metrics": {
744
+ "gpu_monitoring_status": "disabled",
745
+ "gpu_monitoring_reason": "no_gpus_available"
746
+ }
747
+ },
748
+ {
749
+ "scenario_name": "compiled_compile_max-autotune_sdpa_default",
750
+ "metadata": {
751
+ "timestamp": "2025-10-09T16:54:01.650372",
752
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
753
+ "hardware_info": {
754
+ "gpu_name": "unknown",
755
+ "gpu_memory_total_mb": 0,
756
+ "cpu_count": 160,
757
+ "memory_total_mb": 1290069,
758
+ "python_version": "3.12.10",
759
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
760
+ "cuda_version": null
761
+ },
762
+ "config": {
763
+ "name": "compiled",
764
+ "model_id": "meta-llama/Llama-2-7b-hf",
765
+ "variant": "compiled",
766
+ "warmup_iterations": 3,
767
+ "measurement_iterations": 5,
768
+ "num_tokens_to_generate": 100,
769
+ "device": "cuda",
770
+ "torch_dtype": "float16",
771
+ "compile_mode": "max-autotune",
772
+ "compile_options": {},
773
+ "use_cache": true,
774
+ "batch_size": 1,
775
+ "sequence_length": null,
776
+ "attn_implementation": "sdpa",
777
+ "sdpa_backend": null,
778
+ "custom_params": {}
779
+ }
780
+ },
781
+ "measurements": {
782
+ "latency_seconds": {
783
+ "name": "latency_seconds",
784
+ "measurements": [
785
+ 1.422724365234375,
786
+ 1.37926708984375,
787
+ 1.6763681640625,
788
+ 1.5725888671875,
789
+ 1.4589814453125
790
+ ],
791
+ "mean": 1.501985986328125,
792
+ "median": 1.4589814453125,
793
+ "std": 0.10824567739890842,
794
+ "min": 1.37926708984375,
795
+ "max": 1.6763681640625,
796
+ "p25": 1.422724365234375,
797
+ "p75": 1.5725888671875,
798
+ "p90": 1.6348564453125,
799
+ "p95": 1.6556123046875,
800
+ "p99": 1.6722169921875,
801
+ "unit": "seconds"
802
+ },
803
+ "time_to_first_token_seconds": {
804
+ "name": "time_to_first_token_seconds",
805
+ "measurements": [
806
+ 0.014188466072082519,
807
+ 0.01417398738861084,
808
+ 0.01868277168273926,
809
+ 0.017041608810424803,
810
+ 0.015121807098388671
811
+ ],
812
+ "mean": 0.01584172821044922,
813
+ "median": 0.015121807098388671,
814
+ "std": 0.00176318819303063,
815
+ "min": 0.01417398738861084,
816
+ "max": 0.01868277168273926,
817
+ "p25": 0.014188466072082519,
818
+ "p75": 0.017041608810424803,
819
+ "p90": 0.018026306533813475,
820
+ "p95": 0.018354539108276367,
821
+ "p99": 0.01861712516784668,
822
+ "unit": "seconds"
823
+ },
824
+ "tokens_per_second": {
825
+ "name": "tokens_per_second",
826
+ "measurements": [
827
+ 70.2876835763801,
828
+ 72.5022736613896,
829
+ 59.652767300030696,
830
+ 63.58941112106766,
831
+ 68.54096761907822
832
+ ],
833
+ "mean": 66.91462065558926,
834
+ "median": 68.54096761907822,
835
+ "std": 4.669817054930607,
836
+ "min": 59.652767300030696,
837
+ "max": 72.5022736613896,
838
+ "p25": 63.58941112106766,
839
+ "p75": 70.2876835763801,
840
+ "p90": 71.61643762738579,
841
+ "p95": 72.0593556443877,
842
+ "p99": 72.41369005798921,
843
+ "unit": "tokens/sec"
844
+ },
845
+ "time_per_output_token_seconds": {
846
+ "name": "time_per_output_token_seconds",
847
+ "measurements": [
848
+ 0.01422724365234375,
849
+ 0.0137926708984375,
850
+ 0.016763681640625,
851
+ 0.015725888671875,
852
+ 0.014589814453125
853
+ ],
854
+ "mean": 0.015019859863281249,
855
+ "median": 0.014589814453125,
856
+ "std": 0.0010824567739890843,
857
+ "min": 0.0137926708984375,
858
+ "max": 0.016763681640625,
859
+ "p25": 0.01422724365234375,
860
+ "p75": 0.015725888671875,
861
+ "p90": 0.016348564453125,
862
+ "p95": 0.016556123046875,
863
+ "p99": 0.016722169921875,
864
+ "unit": "seconds/token"
865
+ }
866
+ },
867
+ "gpu_metrics": {
868
+ "gpu_monitoring_status": "disabled",
869
+ "gpu_monitoring_reason": "no_gpus_available"
870
+ }
871
+ },
872
+ {
873
+ "scenario_name": "compiled_compile_max-autotune_sdpa_math",
874
+ "metadata": {
875
+ "timestamp": "2025-10-09T16:54:53.025319",
876
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
877
+ "hardware_info": {
878
+ "gpu_name": "unknown",
879
+ "gpu_memory_total_mb": 0,
880
+ "cpu_count": 160,
881
+ "memory_total_mb": 1290069,
882
+ "python_version": "3.12.10",
883
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
884
+ "cuda_version": null
885
+ },
886
+ "config": {
887
+ "name": "compiled",
888
+ "model_id": "meta-llama/Llama-2-7b-hf",
889
+ "variant": "compiled",
890
+ "warmup_iterations": 3,
891
+ "measurement_iterations": 5,
892
+ "num_tokens_to_generate": 100,
893
+ "device": "cuda",
894
+ "torch_dtype": "float16",
895
+ "compile_mode": "max-autotune",
896
+ "compile_options": {},
897
+ "use_cache": true,
898
+ "batch_size": 1,
899
+ "sequence_length": null,
900
+ "attn_implementation": "sdpa",
901
+ "sdpa_backend": "math",
902
+ "custom_params": {}
903
+ }
904
+ },
905
+ "measurements": {
906
+ "latency_seconds": {
907
+ "name": "latency_seconds",
908
+ "measurements": [
909
+ 1.997805908203125,
910
+ 1.7697823486328126,
911
+ 1.8417479248046875,
912
+ 1.948167236328125,
913
+ 1.73647314453125
914
+ ],
915
+ "mean": 1.8587953124999999,
916
+ "median": 1.8417479248046875,
917
+ "std": 0.10048673586768399,
918
+ "min": 1.73647314453125,
919
+ "max": 1.997805908203125,
920
+ "p25": 1.7697823486328126,
921
+ "p75": 1.948167236328125,
922
+ "p90": 1.977950439453125,
923
+ "p95": 1.987878173828125,
924
+ "p99": 1.995820361328125,
925
+ "unit": "seconds"
926
+ },
927
+ "time_to_first_token_seconds": {
928
+ "name": "time_to_first_token_seconds",
929
+ "measurements": [
930
+ 0.020088703155517577,
931
+ 0.020513656616210938,
932
+ 0.025242435455322267,
933
+ 0.020652692794799806,
934
+ 0.020323699951171875
935
+ ],
936
+ "mean": 0.021364237594604492,
937
+ "median": 0.020513656616210938,
938
+ "std": 0.0019483279673867553,
939
+ "min": 0.020088703155517577,
940
+ "max": 0.025242435455322267,
941
+ "p25": 0.020323699951171875,
942
+ "p75": 0.020652692794799806,
943
+ "p90": 0.02340653839111328,
944
+ "p95": 0.024324486923217772,
945
+ "p99": 0.025058845748901368,
946
+ "unit": "seconds"
947
+ },
948
+ "tokens_per_second": {
949
+ "name": "tokens_per_second",
950
+ "measurements": [
951
+ 50.05491253649481,
952
+ 56.504123276657,
953
+ 54.29624687135441,
954
+ 51.33029553893865,
955
+ 57.58799110423004
956
+ ],
957
+ "mean": 53.954713865534984,
958
+ "median": 54.29624687135441,
959
+ "std": 2.8952671238336314,
960
+ "min": 50.05491253649481,
961
+ "max": 57.58799110423004,
962
+ "p25": 51.33029553893865,
963
+ "p75": 56.504123276657,
964
+ "p90": 57.15444397320083,
965
+ "p95": 57.371217538715435,
966
+ "p99": 57.54463639112712,
967
+ "unit": "tokens/sec"
968
+ },
969
+ "time_per_output_token_seconds": {
970
+ "name": "time_per_output_token_seconds",
971
+ "measurements": [
972
+ 0.01997805908203125,
973
+ 0.017697823486328125,
974
+ 0.018417479248046875,
975
+ 0.01948167236328125,
976
+ 0.0173647314453125
977
+ ],
978
+ "mean": 0.018587953125,
979
+ "median": 0.018417479248046875,
980
+ "std": 0.00100486735867684,
981
+ "min": 0.0173647314453125,
982
+ "max": 0.01997805908203125,
983
+ "p25": 0.017697823486328125,
984
+ "p75": 0.01948167236328125,
985
+ "p90": 0.01977950439453125,
986
+ "p95": 0.01987878173828125,
987
+ "p99": 0.01995820361328125,
988
+ "unit": "seconds/token"
989
+ }
990
+ },
991
+ "gpu_metrics": {
992
+ "gpu_monitoring_status": "disabled",
993
+ "gpu_monitoring_reason": "no_gpus_available"
994
+ }
995
+ },
996
+ {
997
+ "scenario_name": "compiled_compile_max-autotune_sdpa_efficient_attention",
998
+ "metadata": {
999
+ "timestamp": "2025-10-09T16:55:57.437430",
1000
+ "commit_id": "c20849bad1df6a557f9a62fa7fb2abf469201296",
1001
+ "hardware_info": {
1002
+ "gpu_name": "unknown",
1003
+ "gpu_memory_total_mb": 0,
1004
+ "cpu_count": 160,
1005
+ "memory_total_mb": 1290069,
1006
+ "python_version": "3.12.10",
1007
+ "torch_version": "2.7.1+rocm6.4.1.git2a215e4a",
1008
+ "cuda_version": null
1009
+ },
1010
+ "config": {
1011
+ "name": "compiled",
1012
+ "model_id": "meta-llama/Llama-2-7b-hf",
1013
+ "variant": "compiled",
1014
+ "warmup_iterations": 3,
1015
+ "measurement_iterations": 5,
1016
+ "num_tokens_to_generate": 100,
1017
+ "device": "cuda",
1018
+ "torch_dtype": "float16",
1019
+ "compile_mode": "max-autotune",
1020
+ "compile_options": {},
1021
+ "use_cache": true,
1022
+ "batch_size": 1,
1023
+ "sequence_length": null,
1024
+ "attn_implementation": "sdpa",
1025
+ "sdpa_backend": "efficient_attention",
1026
+ "custom_params": {}
1027
+ }
1028
+ },
1029
+ "measurements": {
1030
+ "latency_seconds": {
1031
+ "name": "latency_seconds",
1032
+ "measurements": [
1033
+ 1.633405517578125,
1034
+ 1.622461669921875,
1035
+ 1.6361195068359375,
1036
+ 1.6778436279296876,
1037
+ 1.674525390625
1038
+ ],
1039
+ "mean": 1.648871142578125,
1040
+ "median": 1.6361195068359375,
1041
+ "std": 0.022789450923707513,
1042
+ "min": 1.622461669921875,
1043
+ "max": 1.6778436279296876,
1044
+ "p25": 1.633405517578125,
1045
+ "p75": 1.674525390625,
1046
+ "p90": 1.6765163330078126,
1047
+ "p95": 1.67717998046875,
1048
+ "p99": 1.6777108984375,
1049
+ "unit": "seconds"
1050
+ },
1051
+ "time_to_first_token_seconds": {
1052
+ "name": "time_to_first_token_seconds",
1053
+ "measurements": [
1054
+ 0.016919210433959962,
1055
+ 0.017320240020751953,
1056
+ 0.017185123443603515,
1057
+ 0.019705711364746095,
1058
+ 0.01718376350402832
1059
+ ],
1060
+ "mean": 0.01766280975341797,
1061
+ "median": 0.017185123443603515,
1062
+ "std": 0.0010297000031895475,
1063
+ "min": 0.016919210433959962,
1064
+ "max": 0.019705711364746095,
1065
+ "p25": 0.01718376350402832,
1066
+ "p75": 0.017320240020751953,
1067
+ "p90": 0.01875152282714844,
1068
+ "p95": 0.019228617095947267,
1069
+ "p99": 0.01961029251098633,
1070
+ "unit": "seconds"
1071
+ },
1072
+ "tokens_per_second": {
1073
+ "name": "tokens_per_second",
1074
+ "measurements": [
1075
+ 61.22178413372296,
1076
+ 61.63473803656343,
1077
+ 61.12022965448791,
1078
+ 59.60030978774301,
1079
+ 59.718413682981534
1080
+ ],
1081
+ "mean": 60.65909505909976,
1082
+ "median": 61.12022965448791,
1083
+ "std": 0.8351108946069312,
1084
+ "min": 59.60030978774301,
1085
+ "max": 61.63473803656343,
1086
+ "p25": 59.718413682981534,
1087
+ "p75": 61.22178413372296,
1088
+ "p90": 61.46955647542724,
1089
+ "p95": 61.55214725599534,
1090
+ "p99": 61.61821988044981,
1091
+ "unit": "tokens/sec"
1092
+ },
1093
+ "time_per_output_token_seconds": {
1094
+ "name": "time_per_output_token_seconds",
1095
+ "measurements": [
1096
+ 0.01633405517578125,
1097
+ 0.01622461669921875,
1098
+ 0.016361195068359376,
1099
+ 0.016778436279296875,
1100
+ 0.01674525390625
1101
+ ],
1102
+ "mean": 0.01648871142578125,
1103
+ "median": 0.016361195068359376,
1104
+ "std": 0.0002278945092370751,
1105
+ "min": 0.01622461669921875,
1106
+ "max": 0.016778436279296875,
1107
+ "p25": 0.01633405517578125,
1108
+ "p75": 0.01674525390625,
1109
+ "p90": 0.016765163330078124,
1110
+ "p95": 0.0167717998046875,
1111
+ "p99": 0.016777108984375,
1112
+ "unit": "seconds/token"
1113
+ }
1114
+ },
1115
+ "gpu_metrics": {
1116
+ "gpu_monitoring_status": "disabled",
1117
+ "gpu_monitoring_reason": "no_gpus_available"
1118
+ }
1119
+ }
1120
+ ]
1121
+ }