hf-transformers-bot commited on
Commit
de327ee
·
verified ·
1 Parent(s): 4dd4374

Upload benchmark results for run 18104056224

Browse files
2025-09-29/18104056224/benchmark_results/Llama-2-7b-hf/Llama-2-7b-hf_benchmark_20250929_164604.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": "2025-09-29T16:42:51.007154",
8
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
9
+ "hardware_info": {
10
+ "gpu_name": "unknown",
11
+ "gpu_memory_total_mb": 0,
12
+ "cpu_count": 160,
13
+ "memory_total_mb": 1285717,
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
+ 1.823037841796875,
42
+ 1.9389415283203124,
43
+ 1.7958995361328125,
44
+ 2.175377685546875,
45
+ 2.024082763671875
46
+ ],
47
+ "mean": 1.9514678710937503,
48
+ "median": 1.9389415283203124,
49
+ "std": 0.1387524530691676,
50
+ "min": 1.7958995361328125,
51
+ "max": 2.175377685546875,
52
+ "p25": 1.823037841796875,
53
+ "p75": 2.024082763671875,
54
+ "p90": 2.114859716796875,
55
+ "p95": 2.145118701171875,
56
+ "p99": 2.169325888671875,
57
+ "unit": "seconds"
58
+ },
59
+ "time_to_first_token_seconds": {
60
+ "name": "time_to_first_token_seconds",
61
+ "measurements": [
62
+ 0.01879271125793457,
63
+ 0.019078227996826173,
64
+ 0.019155868530273436,
65
+ 0.019397233963012694,
66
+ 0.020517650604248046
67
+ ],
68
+ "mean": 0.019388338470458982,
69
+ "median": 0.019155868530273436,
70
+ "std": 0.0005967256623536848,
71
+ "min": 0.01879271125793457,
72
+ "max": 0.020517650604248046,
73
+ "p25": 0.019078227996826173,
74
+ "p75": 0.019397233963012694,
75
+ "p90": 0.020069483947753906,
76
+ "p95": 0.020293567276000974,
77
+ "p99": 0.020472833938598632,
78
+ "unit": "seconds"
79
+ },
80
+ "tokens_per_second": {
81
+ "name": "tokens_per_second",
82
+ "measurements": [
83
+ 54.85349656891111,
84
+ 51.57453102086534,
85
+ 55.682402043120014,
86
+ 45.96902904005871,
87
+ 49.40509439376415
88
+ ],
89
+ "mean": 51.49691061334386,
90
+ "median": 51.57453102086534,
91
+ "std": 3.5699916666083977,
92
+ "min": 45.96902904005871,
93
+ "max": 55.682402043120014,
94
+ "p25": 49.40509439376415,
95
+ "p75": 54.85349656891111,
96
+ "p90": 55.350839853436455,
97
+ "p95": 55.51662094827823,
98
+ "p99": 55.649245824151656,
99
+ "unit": "tokens/sec"
100
+ },
101
+ "time_per_output_token_seconds": {
102
+ "name": "time_per_output_token_seconds",
103
+ "measurements": [
104
+ 0.01823037841796875,
105
+ 0.019389415283203124,
106
+ 0.017958995361328126,
107
+ 0.02175377685546875,
108
+ 0.02024082763671875
109
+ ],
110
+ "mean": 0.0195146787109375,
111
+ "median": 0.019389415283203124,
112
+ "std": 0.0013875245306916762,
113
+ "min": 0.017958995361328126,
114
+ "max": 0.02175377685546875,
115
+ "p25": 0.01823037841796875,
116
+ "p75": 0.02024082763671875,
117
+ "p90": 0.02114859716796875,
118
+ "p95": 0.02145118701171875,
119
+ "p99": 0.02169325888671875,
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": "2025-09-29T16:43:12.536025",
132
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
133
+ "hardware_info": {
134
+ "gpu_name": "unknown",
135
+ "gpu_memory_total_mb": 0,
136
+ "cpu_count": 160,
137
+ "memory_total_mb": 1285717,
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
+ 3.38871533203125,
166
+ 3.43227099609375,
167
+ 3.39078564453125,
168
+ 3.420824462890625,
169
+ 3.448164306640625
170
+ ],
171
+ "mean": 3.4161521484375,
172
+ "median": 3.420824462890625,
173
+ "std": 0.0232493632595487,
174
+ "min": 3.38871533203125,
175
+ "max": 3.448164306640625,
176
+ "p25": 3.39078564453125,
177
+ "p75": 3.43227099609375,
178
+ "p90": 3.441806982421875,
179
+ "p95": 3.44498564453125,
180
+ "p99": 3.44752857421875,
181
+ "unit": "seconds"
182
+ },
183
+ "time_to_first_token_seconds": {
184
+ "name": "time_to_first_token_seconds",
185
+ "measurements": [
186
+ 0.018727634429931642,
187
+ 0.01911126708984375,
188
+ 0.019572864532470702,
189
+ 0.018379840850830078,
190
+ 0.018644676208496094
191
+ ],
192
+ "mean": 0.018887256622314455,
193
+ "median": 0.018727634429931642,
194
+ "std": 0.00041521901754423016,
195
+ "min": 0.018379840850830078,
196
+ "max": 0.019572864532470702,
197
+ "p25": 0.018644676208496094,
198
+ "p75": 0.01911126708984375,
199
+ "p90": 0.01938822555541992,
200
+ "p95": 0.01948054504394531,
201
+ "p99": 0.019554400634765625,
202
+ "unit": "seconds"
203
+ },
204
+ "tokens_per_second": {
205
+ "name": "tokens_per_second",
206
+ "measurements": [
207
+ 29.509708016712754,
208
+ 29.135228574261614,
209
+ 29.49169026985905,
210
+ 29.232718920484793,
211
+ 29.000938211504494
212
+ ],
213
+ "mean": 29.274056798564537,
214
+ "median": 29.232718920484793,
215
+ "std": 0.19923437608372238,
216
+ "min": 29.000938211504494,
217
+ "max": 29.509708016712754,
218
+ "p25": 29.135228574261614,
219
+ "p75": 29.49169026985905,
220
+ "p90": 29.502500917971272,
221
+ "p95": 29.506104467342013,
222
+ "p99": 29.508987306838605,
223
+ "unit": "tokens/sec"
224
+ },
225
+ "time_per_output_token_seconds": {
226
+ "name": "time_per_output_token_seconds",
227
+ "measurements": [
228
+ 0.0338871533203125,
229
+ 0.0343227099609375,
230
+ 0.033907856445312495,
231
+ 0.03420824462890625,
232
+ 0.03448164306640625
233
+ ],
234
+ "mean": 0.034161521484375,
235
+ "median": 0.03420824462890625,
236
+ "std": 0.00023249363259548667,
237
+ "min": 0.0338871533203125,
238
+ "max": 0.03448164306640625,
239
+ "p25": 0.033907856445312495,
240
+ "p75": 0.0343227099609375,
241
+ "p90": 0.03441806982421875,
242
+ "p95": 0.034449856445312496,
243
+ "p99": 0.0344752857421875,
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": "2025-09-29T16:43:59.416382",
256
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
257
+ "hardware_info": {
258
+ "gpu_name": "unknown",
259
+ "gpu_memory_total_mb": 0,
260
+ "cpu_count": 160,
261
+ "memory_total_mb": 1285717,
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
+ 1.4153052978515626,
290
+ 1.4144854736328125,
291
+ 1.414724609375,
292
+ 1.435518310546875,
293
+ 1.439624267578125
294
+ ],
295
+ "mean": 1.423931591796875,
296
+ "median": 1.4153052978515626,
297
+ "std": 0.011215371108296946,
298
+ "min": 1.4144854736328125,
299
+ "max": 1.439624267578125,
300
+ "p25": 1.414724609375,
301
+ "p75": 1.435518310546875,
302
+ "p90": 1.437981884765625,
303
+ "p95": 1.438803076171875,
304
+ "p99": 1.439460029296875,
305
+ "unit": "seconds"
306
+ },
307
+ "time_to_first_token_seconds": {
308
+ "name": "time_to_first_token_seconds",
309
+ "measurements": [
310
+ 0.015208359718322753,
311
+ 0.015403918266296387,
312
+ 0.015227800369262695,
313
+ 0.015441157341003419,
314
+ 0.01562551498413086
315
+ ],
316
+ "mean": 0.015381350135803223,
317
+ "median": 0.015403918266296387,
318
+ "std": 0.00015310513434329423,
319
+ "min": 0.015208359718322753,
320
+ "max": 0.01562551498413086,
321
+ "p25": 0.015227800369262695,
322
+ "p75": 0.015441157341003419,
323
+ "p90": 0.015551771926879884,
324
+ "p95": 0.015588643455505372,
325
+ "p99": 0.015618140678405763,
326
+ "unit": "seconds"
327
+ },
328
+ "tokens_per_second": {
329
+ "name": "tokens_per_second",
330
+ "measurements": [
331
+ 70.65613345177205,
332
+ 70.6970851691893,
333
+ 70.68513499894387,
334
+ 69.66125006228866,
335
+ 69.46256898560738
336
+ ],
337
+ "mean": 70.23243453356025,
338
+ "median": 70.65613345177205,
339
+ "std": 0.5512355866577288,
340
+ "min": 69.46256898560738,
341
+ "max": 70.6970851691893,
342
+ "p25": 69.66125006228866,
343
+ "p75": 70.68513499894387,
344
+ "p90": 70.69230510109112,
345
+ "p95": 70.69469513514021,
346
+ "p99": 70.69660716237948,
347
+ "unit": "tokens/sec"
348
+ },
349
+ "time_per_output_token_seconds": {
350
+ "name": "time_per_output_token_seconds",
351
+ "measurements": [
352
+ 0.014153052978515625,
353
+ 0.014144854736328125,
354
+ 0.014147246093750001,
355
+ 0.014355183105468749,
356
+ 0.01439624267578125
357
+ ],
358
+ "mean": 0.01423931591796875,
359
+ "median": 0.014153052978515625,
360
+ "std": 0.0001121537110829694,
361
+ "min": 0.014144854736328125,
362
+ "max": 0.01439624267578125,
363
+ "p25": 0.014147246093750001,
364
+ "p75": 0.014355183105468749,
365
+ "p90": 0.01437981884765625,
366
+ "p95": 0.01438803076171875,
367
+ "p99": 0.01439460029296875,
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": "2025-09-29T16:44:17.492664",
380
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
381
+ "hardware_info": {
382
+ "gpu_name": "unknown",
383
+ "gpu_memory_total_mb": 0,
384
+ "cpu_count": 160,
385
+ "memory_total_mb": 1285717,
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
+ 1.4897742919921875,
414
+ 1.74798388671875,
415
+ 1.7545491943359375,
416
+ 1.7508719482421875,
417
+ 1.332464111328125
418
+ ],
419
+ "mean": 1.6151286865234376,
420
+ "median": 1.74798388671875,
421
+ "std": 0.17385499798298565,
422
+ "min": 1.332464111328125,
423
+ "max": 1.7545491943359375,
424
+ "p25": 1.4897742919921875,
425
+ "p75": 1.7508719482421875,
426
+ "p90": 1.7530782958984374,
427
+ "p95": 1.7538137451171876,
428
+ "p99": 1.7544021044921876,
429
+ "unit": "seconds"
430
+ },
431
+ "time_to_first_token_seconds": {
432
+ "name": "time_to_first_token_seconds",
433
+ "measurements": [
434
+ 0.016038429260253905,
435
+ 0.01927214813232422,
436
+ 0.01959998893737793,
437
+ 0.020609376907348632,
438
+ 0.016079830169677733
439
+ ],
440
+ "mean": 0.018319954681396484,
441
+ "median": 0.01927214813232422,
442
+ "std": 0.0018979004330302645,
443
+ "min": 0.016038429260253905,
444
+ "max": 0.020609376907348632,
445
+ "p25": 0.016079830169677733,
446
+ "p75": 0.01959998893737793,
447
+ "p90": 0.020205621719360353,
448
+ "p95": 0.02040749931335449,
449
+ "p99": 0.020569001388549803,
450
+ "unit": "seconds"
451
+ },
452
+ "tokens_per_second": {
453
+ "name": "tokens_per_second",
454
+ "measurements": [
455
+ 67.12426206944133,
456
+ 57.20876534377915,
457
+ 56.994697169404844,
458
+ 57.11439954269437,
459
+ 75.04892563322073
460
+ ],
461
+ "mean": 62.69820995170808,
462
+ "median": 57.20876534377915,
463
+ "std": 7.293464389982031,
464
+ "min": 56.994697169404844,
465
+ "max": 75.04892563322073,
466
+ "p25": 57.11439954269437,
467
+ "p75": 67.12426206944133,
468
+ "p90": 71.87906020770897,
469
+ "p95": 73.46399292046485,
470
+ "p99": 74.73193909066956,
471
+ "unit": "tokens/sec"
472
+ },
473
+ "time_per_output_token_seconds": {
474
+ "name": "time_per_output_token_seconds",
475
+ "measurements": [
476
+ 0.014897742919921875,
477
+ 0.0174798388671875,
478
+ 0.017545491943359373,
479
+ 0.017508719482421876,
480
+ 0.01332464111328125
481
+ ],
482
+ "mean": 0.016151286865234378,
483
+ "median": 0.0174798388671875,
484
+ "std": 0.0017385499798298562,
485
+ "min": 0.01332464111328125,
486
+ "max": 0.017545491943359373,
487
+ "p25": 0.014897742919921875,
488
+ "p75": 0.017508719482421876,
489
+ "p90": 0.017530782958984373,
490
+ "p95": 0.017538137451171873,
491
+ "p99": 0.01754402104492187,
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": "2025-09-29T16:44:37.707828",
504
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
505
+ "hardware_info": {
506
+ "gpu_name": "unknown",
507
+ "gpu_memory_total_mb": 0,
508
+ "cpu_count": 160,
509
+ "memory_total_mb": 1285717,
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
+ 1.44450341796875,
538
+ 1.4614754638671874,
539
+ 1.4575396728515626,
540
+ 1.5042498779296876,
541
+ 1.4562506103515624
542
+ ],
543
+ "mean": 1.46480380859375,
544
+ "median": 1.4575396728515626,
545
+ "std": 0.020518776729878934,
546
+ "min": 1.44450341796875,
547
+ "max": 1.5042498779296876,
548
+ "p25": 1.4562506103515624,
549
+ "p75": 1.4614754638671874,
550
+ "p90": 1.4871401123046875,
551
+ "p95": 1.4956949951171876,
552
+ "p99": 1.5025389013671875,
553
+ "unit": "seconds"
554
+ },
555
+ "time_to_first_token_seconds": {
556
+ "name": "time_to_first_token_seconds",
557
+ "measurements": [
558
+ 0.015339638710021973,
559
+ 0.015895352363586426,
560
+ 0.015451278686523438,
561
+ 0.01576219367980957,
562
+ 0.01607419204711914
563
+ ],
564
+ "mean": 0.015704531097412107,
565
+ "median": 0.01576219367980957,
566
+ "std": 0.0002733752676331987,
567
+ "min": 0.015339638710021973,
568
+ "max": 0.01607419204711914,
569
+ "p25": 0.015451278686523438,
570
+ "p75": 0.015895352363586426,
571
+ "p90": 0.016002656173706054,
572
+ "p95": 0.016038424110412598,
573
+ "p99": 0.016067038459777833,
574
+ "unit": "seconds"
575
+ },
576
+ "tokens_per_second": {
577
+ "name": "tokens_per_second",
578
+ "measurements": [
579
+ 69.22794280446858,
580
+ 68.42400195716701,
581
+ 68.60876713177748,
582
+ 66.47831684562334,
583
+ 68.66949911585506
584
+ ],
585
+ "mean": 68.28170557097829,
586
+ "median": 68.60876713177748,
587
+ "std": 0.9407632174254668,
588
+ "min": 66.47831684562334,
589
+ "max": 69.22794280446858,
590
+ "p25": 68.42400195716701,
591
+ "p75": 68.66949911585506,
592
+ "p90": 69.00456532902318,
593
+ "p95": 69.11625406674588,
594
+ "p99": 69.20560505692404,
595
+ "unit": "tokens/sec"
596
+ },
597
+ "time_per_output_token_seconds": {
598
+ "name": "time_per_output_token_seconds",
599
+ "measurements": [
600
+ 0.0144450341796875,
601
+ 0.014614754638671875,
602
+ 0.014575396728515625,
603
+ 0.015042498779296875,
604
+ 0.014562506103515624
605
+ ],
606
+ "mean": 0.0146480380859375,
607
+ "median": 0.014575396728515625,
608
+ "std": 0.000205187767298789,
609
+ "min": 0.0144450341796875,
610
+ "max": 0.015042498779296875,
611
+ "p25": 0.014562506103515624,
612
+ "p75": 0.014614754638671875,
613
+ "p90": 0.014871401123046874,
614
+ "p95": 0.014956949951171875,
615
+ "p99": 0.015025389013671875,
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": "2025-09-29T16:44:56.213783",
628
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
629
+ "hardware_info": {
630
+ "gpu_name": "unknown",
631
+ "gpu_memory_total_mb": 0,
632
+ "cpu_count": 160,
633
+ "memory_total_mb": 1285717,
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
+ 1.4119998779296874,
662
+ 1.1978817138671876,
663
+ 1.2877467041015624,
664
+ 1.2060032958984375,
665
+ 1.2066937255859376
666
+ ],
667
+ "mean": 1.2620650634765624,
668
+ "median": 1.2066937255859376,
669
+ "std": 0.08181491851874614,
670
+ "min": 1.1978817138671876,
671
+ "max": 1.4119998779296874,
672
+ "p25": 1.2060032958984375,
673
+ "p75": 1.2877467041015624,
674
+ "p90": 1.3622986083984374,
675
+ "p95": 1.3871492431640624,
676
+ "p99": 1.4070297509765624,
677
+ "unit": "seconds"
678
+ },
679
+ "time_to_first_token_seconds": {
680
+ "name": "time_to_first_token_seconds",
681
+ "measurements": [
682
+ 0.01496312427520752,
683
+ 0.012057441711425781,
684
+ 0.012129040718078613,
685
+ 0.013240386009216308,
686
+ 0.013823939323425292
687
+ ],
688
+ "mean": 0.013242786407470702,
689
+ "median": 0.013240386009216308,
690
+ "std": 0.001090209003988212,
691
+ "min": 0.012057441711425781,
692
+ "max": 0.01496312427520752,
693
+ "p25": 0.012129040718078613,
694
+ "p75": 0.013823939323425292,
695
+ "p90": 0.014507450294494628,
696
+ "p95": 0.014735287284851073,
697
+ "p99": 0.01491755687713623,
698
+ "unit": "seconds"
699
+ },
700
+ "tokens_per_second": {
701
+ "name": "tokens_per_second",
702
+ "measurements": [
703
+ 70.82153586771035,
704
+ 83.48069666842521,
705
+ 77.6550230581007,
706
+ 82.91851302570687,
707
+ 82.8710698329377
708
+ ],
709
+ "mean": 79.54936769057618,
710
+ "median": 82.8710698329377,
711
+ "std": 4.849820261797249,
712
+ "min": 70.82153586771035,
713
+ "max": 83.48069666842521,
714
+ "p25": 77.6550230581007,
715
+ "p75": 82.91851302570687,
716
+ "p90": 83.25582321133787,
717
+ "p95": 83.36825993988154,
718
+ "p99": 83.45820932271647,
719
+ "unit": "tokens/sec"
720
+ },
721
+ "time_per_output_token_seconds": {
722
+ "name": "time_per_output_token_seconds",
723
+ "measurements": [
724
+ 0.014119998779296873,
725
+ 0.011978817138671875,
726
+ 0.012877467041015625,
727
+ 0.012060032958984375,
728
+ 0.012066937255859375
729
+ ],
730
+ "mean": 0.012620650634765623,
731
+ "median": 0.012066937255859375,
732
+ "std": 0.0008181491851874615,
733
+ "min": 0.011978817138671875,
734
+ "max": 0.014119998779296873,
735
+ "p25": 0.012060032958984375,
736
+ "p75": 0.012877467041015625,
737
+ "p90": 0.013622986083984373,
738
+ "p95": 0.013871492431640623,
739
+ "p99": 0.014070297509765623,
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": "2025-09-29T16:45:11.731448",
752
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
753
+ "hardware_info": {
754
+ "gpu_name": "unknown",
755
+ "gpu_memory_total_mb": 0,
756
+ "cpu_count": 160,
757
+ "memory_total_mb": 1285717,
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
+ 1.4372919921875,
786
+ 1.2798974609375,
787
+ 1.3733223876953125,
788
+ 1.3982735595703124,
789
+ 1.449629638671875
790
+ ],
791
+ "mean": 1.3876830078125,
792
+ "median": 1.3982735595703124,
793
+ "std": 0.060389647651521466,
794
+ "min": 1.2798974609375,
795
+ "max": 1.449629638671875,
796
+ "p25": 1.3733223876953125,
797
+ "p75": 1.4372919921875,
798
+ "p90": 1.444694580078125,
799
+ "p95": 1.447162109375,
800
+ "p99": 1.4491361328125,
801
+ "unit": "seconds"
802
+ },
803
+ "time_to_first_token_seconds": {
804
+ "name": "time_to_first_token_seconds",
805
+ "measurements": [
806
+ 0.016329185485839844,
807
+ 0.013605941772460937,
808
+ 0.013594621658325196,
809
+ 0.015121521949768067,
810
+ 0.01518208122253418
811
+ ],
812
+ "mean": 0.014766670417785645,
813
+ "median": 0.015121521949768067,
814
+ "std": 0.0010450768743952403,
815
+ "min": 0.013594621658325196,
816
+ "max": 0.016329185485839844,
817
+ "p25": 0.013605941772460937,
818
+ "p75": 0.01518208122253418,
819
+ "p90": 0.01587034378051758,
820
+ "p95": 0.01609976463317871,
821
+ "p99": 0.016283301315307617,
822
+ "unit": "seconds"
823
+ },
824
+ "tokens_per_second": {
825
+ "name": "tokens_per_second",
826
+ "measurements": [
827
+ 69.5752850106707,
828
+ 78.13125898910053,
829
+ 72.81611433409921,
830
+ 71.51676388040254,
831
+ 68.9831370249978
832
+ ],
833
+ "mean": 72.20451184785415,
834
+ "median": 71.51676388040254,
835
+ "std": 3.263846689257496,
836
+ "min": 68.9831370249978,
837
+ "max": 78.13125898910053,
838
+ "p25": 69.5752850106707,
839
+ "p75": 72.81611433409921,
840
+ "p90": 76.00520112710001,
841
+ "p95": 77.06823005810027,
842
+ "p99": 77.91865320290047,
843
+ "unit": "tokens/sec"
844
+ },
845
+ "time_per_output_token_seconds": {
846
+ "name": "time_per_output_token_seconds",
847
+ "measurements": [
848
+ 0.014372919921875,
849
+ 0.012798974609375,
850
+ 0.013733223876953124,
851
+ 0.013982735595703124,
852
+ 0.01449629638671875
853
+ ],
854
+ "mean": 0.013876830078125,
855
+ "median": 0.013982735595703124,
856
+ "std": 0.0006038964765152147,
857
+ "min": 0.012798974609375,
858
+ "max": 0.01449629638671875,
859
+ "p25": 0.013733223876953124,
860
+ "p75": 0.014372919921875,
861
+ "p90": 0.01444694580078125,
862
+ "p95": 0.01447162109375,
863
+ "p99": 0.014491361328124999,
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": "2025-09-29T16:45:28.154789",
876
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
877
+ "hardware_info": {
878
+ "gpu_name": "unknown",
879
+ "gpu_memory_total_mb": 0,
880
+ "cpu_count": 160,
881
+ "memory_total_mb": 1285717,
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
+ 1.70909521484375,
910
+ 1.718250732421875,
911
+ 1.714868896484375,
912
+ 1.7014049072265625,
913
+ 1.5307169189453125
914
+ ],
915
+ "mean": 1.6748673339843747,
916
+ "median": 1.70909521484375,
917
+ "std": 0.07230128220172903,
918
+ "min": 1.5307169189453125,
919
+ "max": 1.718250732421875,
920
+ "p25": 1.7014049072265625,
921
+ "p75": 1.714868896484375,
922
+ "p90": 1.716897998046875,
923
+ "p95": 1.717574365234375,
924
+ "p99": 1.718115458984375,
925
+ "unit": "seconds"
926
+ },
927
+ "time_to_first_token_seconds": {
928
+ "name": "time_to_first_token_seconds",
929
+ "measurements": [
930
+ 0.018086523056030274,
931
+ 0.01824812126159668,
932
+ 0.018147043228149414,
933
+ 0.018757755279541016,
934
+ 0.01683177947998047
935
+ ],
936
+ "mean": 0.01801424446105957,
937
+ "median": 0.018147043228149414,
938
+ "std": 0.0006369599413367129,
939
+ "min": 0.01683177947998047,
940
+ "max": 0.018757755279541016,
941
+ "p25": 0.018086523056030274,
942
+ "p75": 0.01824812126159668,
943
+ "p90": 0.01855390167236328,
944
+ "p95": 0.01865582847595215,
945
+ "p99": 0.018737369918823243,
946
+ "unit": "seconds"
947
+ },
948
+ "tokens_per_second": {
949
+ "name": "tokens_per_second",
950
+ "measurements": [
951
+ 58.51049089101936,
952
+ 58.19872391909298,
953
+ 58.31349568763442,
954
+ 58.77495684610941,
955
+ 65.32886568530355
956
+ ],
957
+ "mean": 59.825306605831955,
958
+ "median": 58.51049089101936,
959
+ "std": 2.7587127799784414,
960
+ "min": 58.19872391909298,
961
+ "max": 65.32886568530355,
962
+ "p25": 58.31349568763442,
963
+ "p75": 58.77495684610941,
964
+ "p90": 62.707302149625896,
965
+ "p95": 64.01808391746472,
966
+ "p99": 65.0667093317358,
967
+ "unit": "tokens/sec"
968
+ },
969
+ "time_per_output_token_seconds": {
970
+ "name": "time_per_output_token_seconds",
971
+ "measurements": [
972
+ 0.017090952148437498,
973
+ 0.01718250732421875,
974
+ 0.01714868896484375,
975
+ 0.017014049072265624,
976
+ 0.015307169189453124
977
+ ],
978
+ "mean": 0.01674867333984375,
979
+ "median": 0.017090952148437498,
980
+ "std": 0.0007230128220172902,
981
+ "min": 0.015307169189453124,
982
+ "max": 0.01718250732421875,
983
+ "p25": 0.017014049072265624,
984
+ "p75": 0.01714868896484375,
985
+ "p90": 0.01716897998046875,
986
+ "p95": 0.01717574365234375,
987
+ "p99": 0.017181154589843747,
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": "2025-09-29T16:45:51.990082",
1000
+ "commit_id": "8635d8e796c77ee94f64c170168f71138380274b",
1001
+ "hardware_info": {
1002
+ "gpu_name": "unknown",
1003
+ "gpu_memory_total_mb": 0,
1004
+ "cpu_count": 160,
1005
+ "memory_total_mb": 1285717,
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
+ 1.4626751708984376,
1034
+ 1.4810311279296875,
1035
+ 1.4736494140625,
1036
+ 1.4933232421875,
1037
+ 1.4591546630859376
1038
+ ],
1039
+ "mean": 1.4739667236328127,
1040
+ "median": 1.4736494140625,
1041
+ "std": 0.012422379842755887,
1042
+ "min": 1.4591546630859376,
1043
+ "max": 1.4933232421875,
1044
+ "p25": 1.4626751708984376,
1045
+ "p75": 1.4810311279296875,
1046
+ "p90": 1.488406396484375,
1047
+ "p95": 1.4908648193359375,
1048
+ "p99": 1.4928315576171876,
1049
+ "unit": "seconds"
1050
+ },
1051
+ "time_to_first_token_seconds": {
1052
+ "name": "time_to_first_token_seconds",
1053
+ "measurements": [
1054
+ 0.014981563568115235,
1055
+ 0.015441958427429198,
1056
+ 0.015426198005676269,
1057
+ 0.015393279075622558,
1058
+ 0.015228960990905762
1059
+ ],
1060
+ "mean": 0.015294392013549804,
1061
+ "median": 0.015393279075622558,
1062
+ "std": 0.00017382184679050054,
1063
+ "min": 0.014981563568115235,
1064
+ "max": 0.015441958427429198,
1065
+ "p25": 0.015228960990905762,
1066
+ "p75": 0.015426198005676269,
1067
+ "p90": 0.015435654258728027,
1068
+ "p95": 0.015438806343078612,
1069
+ "p99": 0.01544132801055908,
1070
+ "unit": "seconds"
1071
+ },
1072
+ "tokens_per_second": {
1073
+ "name": "tokens_per_second",
1074
+ "measurements": [
1075
+ 68.36787961511354,
1076
+ 67.52052547321445,
1077
+ 67.85874513010788,
1078
+ 66.96473822607531,
1079
+ 68.5328310492542
1080
+ ],
1081
+ "mean": 67.84894389875308,
1082
+ "median": 67.85874513010788,
1083
+ "std": 0.5703981609893162,
1084
+ "min": 66.96473822607531,
1085
+ "max": 68.5328310492542,
1086
+ "p25": 67.52052547321445,
1087
+ "p75": 68.36787961511354,
1088
+ "p90": 68.46685047559794,
1089
+ "p95": 68.49984076242606,
1090
+ "p99": 68.52623299188858,
1091
+ "unit": "tokens/sec"
1092
+ },
1093
+ "time_per_output_token_seconds": {
1094
+ "name": "time_per_output_token_seconds",
1095
+ "measurements": [
1096
+ 0.014626751708984375,
1097
+ 0.014810311279296875,
1098
+ 0.014736494140625,
1099
+ 0.014933232421875,
1100
+ 0.014591546630859376
1101
+ ],
1102
+ "mean": 0.014739667236328125,
1103
+ "median": 0.014736494140625,
1104
+ "std": 0.00012422379842755887,
1105
+ "min": 0.014591546630859376,
1106
+ "max": 0.014933232421875,
1107
+ "p25": 0.014626751708984375,
1108
+ "p75": 0.014810311279296875,
1109
+ "p90": 0.01488406396484375,
1110
+ "p95": 0.014908648193359375,
1111
+ "p99": 0.014928315576171875,
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
+ }