107 lines
16 KiB
Plaintext
107 lines
16 KiB
Plaintext
test-gpt2-kernels - TF32 Mode is turned OFF.
|
|
test-gpt2 - [State]
|
|
Batch Size: 4
|
|
Sequence Length: 64
|
|
GPT2 - Allocated 221 MiB for activations
|
|
test-gpt2 - Printing 10 Expected vs Actual Logits:
|
|
-43.431618, -43.431740
|
|
-39.836346, -39.836452
|
|
-43.065910, -43.066021
|
|
-42.828045, -42.828144
|
|
-43.529541, -43.529644
|
|
-44.318398, -44.318497
|
|
-41.227425, -41.227528
|
|
-41.270760, -41.270863
|
|
-42.541393, -42.541523
|
|
-42.394997, -42.395119
|
|
---------------- Logits OK! (max err 0.000043, RMSE 0.000002) ----------------
|
|
_________________________ All tests passed! _________________________
|
|
Collecting data...
|
|
Generating '/tmp/nsys-report-a416.qdstrm'
|
|
|
|
[1/8] [0% ] report1.nsys-rep
|
|
[1/8] [0% ] report1.nsys-rep
|
|
[1/8] [====25% ] report1.nsys-rep
|
|
[1/8] [==========47% ] report1.nsys-rep
|
|
[1/8] [==========48% ] report1.nsys-rep
|
|
[1/8] [==========49% ] report1.nsys-rep
|
|
[1/8] [===========50% ] report1.nsys-rep
|
|
[1/8] [======34% ] report1.nsys-rep
|
|
[1/8] [======35% ] report1.nsys-rep
|
|
[1/8] [===========50% ] report1.nsys-rep
|
|
[1/8] [===========51% ] report1.nsys-rep
|
|
[1/8] [===========52% ] report1.nsys-rep
|
|
[1/8] [===========53% ] report1.nsys-rep
|
|
[1/8] [==================77% ] report1.nsys-rep
|
|
[1/8] [======================91% ] report1.nsys-rep
|
|
[1/8] [========================100%] report1.nsys-rep
|
|
[1/8] [========================100%] report1.nsys-rep
|
|
|
|
[2/8] [0% ] report1.sqlite
|
|
[2/8] [1% ] report1.sqlite
|
|
[2/8] [2% ] report1.sqlite
|
|
[2/8] [3% ] report1.sqlite
|
|
[2/8] [4% ] report1.sqlite
|
|
[2/8] [5% ] report1.sqlite
|
|
[2/8] [6% ] report1.sqlite
|
|
[2/8] [7% ] report1.sqlite
|
|
[2/8] [8% ] report1.sqlite
|
|
[2/8] [9% ] report1.sqlite
|
|
[2/8] [10% ] report1.sqlite
|
|
[2/8] [11% ] report1.sqlite
|
|
[2/8] [12% ] report1.sqlite
|
|
[2/8] [13% ] report1.sqlite
|
|
[2/8] [14% ] report1.sqlite
|
|
[2/8] [=15% ] report1.sqlite
|
|
[2/8] [=16% ] report1.sqlite
|
|
[2/8] [=17% ] report1.sqlite
|
|
[2/8] [==18% ] report1.sqlite
|
|
[2/8] [==19% ] report1.sqlite
|
|
[2/8] [==20% ] report1.sqlite
|
|
[2/8] [==21% ] report1.sqlite
|
|
[2/8] [===22% ] report1.sqlite
|
|
[2/8] [===23% ] report1.sqlite
|
|
[2/8] [===24% ] report1.sqlite
|
|
[2/8] [====25% ] report1.sqlite
|
|
[2/8] [====26% ] report1.sqlite
|
|
[2/8] [====27% ] report1.sqlite
|
|
[2/8] [====28% ] report1.sqlite
|
|
[2/8] [=====29% ] report1.sqlite
|
|
[2/8] [=====30% ] report1.sqlite
|
|
[2/8] [=====31% ] report1.sqlite
|
|
[2/8] [=====32% ] report1.sqlite
|
|
[2/8] [======33% ] report1.sqlite
|
|
[2/8] [======34% ] report1.sqlite
|
|
[2/8] [======35% ] report1.sqlite
|
|
[2/8] [=======36% ] report1.sqlite
|
|
[2/8] [=======37% ] report1.sqlite
|
|
[2/8] [=======38% ] report1.sqlite
|
|
[2/8] [=======39% ] report1.sqlite
|
|
[2/8] [========40% ] report1.sqlite
|
|
[2/8] [========41% ] report1.sqlite
|
|
[2/8] [========42% ] report1.sqlite
|
|
[2/8] [=========43% ] report1.sqlite
|
|
[2/8] [=========44% ] report1.sqlite
|
|
[2/8] [=========45% ] report1.sqlite
|
|
[2/8] [=========46% ] report1.sqlite
|
|
[2/8] [==========47% ] report1.sqlite
|
|
[2/8] [==========48% ] report1.sqlite
|
|
[2/8] [==========49% ] report1.sqlite
|
|
[2/8] [===========50% ] report1.sqlite
|
|
[2/8] [===========51% ] report1.sqlite
|
|
[2/8] [===========52% ] report1.sqlite
|
|
[2/8] [===========53% ] report1.sqlite
|
|
[2/8] [============54% ] report1.sqlite
|
|
[2/8] [============55% ] report1.sqlite
|
|
[2/8] [============56% ] report1.sqlite
|
|
[2/8] [============57% ] report1.sqlite
|
|
[2/8] [=============58% ] report1.sqlite
|
|
[2/8] [=============59% ] report1.sqlite
|
|
[2/8] [=============60% ] report1.sqlite
|
|
[2/8] [==============61% ] report1.sqlite
|
|
[2/8] [==============62% ] report1.sqlite
|
|
[2/8] [==============63% ] report1.sqlite
|
|
[2/8] [==============64% ] report1.sqlite
|
|
[2/8] [===============65% ] report1.sqlite
|
|
[2/8] [===============66% ] report1.sqlite
|
|
[2/8] [===============67% ] report1.sqlite
|
|
[2/8] [================68% ] report1.sqlite
|
|
[2/8] [================69% ] report1.sqlite
|
|
[2/8] [================70% ] report1.sqlite
|
|
[2/8] [================71% ] report1.sqlite
|
|
[2/8] [=================72% ] report1.sqlite
|
|
[2/8] [=================73% ] report1.sqlite
|
|
[2/8] [=================74% ] report1.sqlite
|
|
[2/8] [==================75% ] report1.sqlite
|
|
[2/8] [==================76% ] report1.sqlite
|
|
[2/8] [==================77% ] report1.sqlite
|
|
[2/8] [==================78% ] report1.sqlite
|
|
[2/8] [===================79% ] report1.sqlite
|
|
[2/8] [===================80% ] report1.sqlite
|
|
[2/8] [===================81% ] report1.sqlite
|
|
[2/8] [===================82% ] report1.sqlite
|
|
[2/8] [====================83% ] report1.sqlite
|
|
[2/8] [====================84% ] report1.sqlite
|
|
[2/8] [====================85% ] report1.sqlite
|
|
[2/8] [=====================86% ] report1.sqlite
|
|
[2/8] [=====================87% ] report1.sqlite
|
|
[2/8] [=====================88% ] report1.sqlite
|
|
[2/8] [=====================89% ] report1.sqlite
|
|
[2/8] [======================90% ] report1.sqlite
|
|
[2/8] [======================91% ] report1.sqlite
|
|
[2/8] [======================92% ] report1.sqlite
|
|
[2/8] [=======================93% ] report1.sqlite
|
|
[2/8] [=======================94% ] report1.sqlite
|
|
[2/8] [=======================95% ] report1.sqlite
|
|
[2/8] [=======================96% ] report1.sqlite
|
|
[2/8] [========================97% ] report1.sqlite
|
|
[2/8] [========================98% ] report1.sqlite
|
|
[2/8] [========================99% ] report1.sqlite
|
|
[2/8] [========================100%] report1.sqlite
|
|
[2/8] [========================100%] report1.sqlite
|
|
[3/8] Executing 'nvtx_sum' stats report
|
|
[4/8] Executing 'osrt_sum' stats report
|
|
|
|
Time (%) Total Time (ns) Num Calls Avg (ns) Med (ns) Min (ns) Max (ns) StdDev (ns) Name
|
|
-------- --------------- --------- ---------- ---------- -------- --------- ----------- ----------------------
|
|
72.0 705681437 17 41510672.8 22601905.0 1263 105099462 46154890.8 poll
|
|
12.0 117682190 590 199461.3 23083.0 1022 21896375 998390.5 ioctl
|
|
8.9 86890629 6 14481771.5 81337.5 7083 79381817 31927343.1 fread
|
|
4.5 44498140 38 1171003.7 6392.0 3587 43521912 7056305.1 fopen
|
|
1.3 12418008 1 12418008.0 12418008.0 12418008 12418008 0.0 pthread_cond_wait
|
|
0.7 7195204 1 7195204.0 7195204.0 7195204 7195204 0.0 pthread_cond_broadcast
|
|
0.2 2355779 25 94231.2 8807.0 6823 1637274 324132.7 mmap64
|
|
0.1 1272540 10 127254.0 31053.0 21640 545090 194577.0 write
|
|
0.1 726199 32 22693.7 2144.0 1343 361726 80378.1 fclose
|
|
0.1 601015 17 35353.8 7755.0 1883 391553 94165.5 mmap
|
|
0.0 368309 2 184154.5 184154.5 168145 200164 22640.9 pthread_create
|
|
0.0 316921 55 5762.2 5831.0 2354 10750 2150.6 open64
|
|
0.0 78106 1 78106.0 78106.0 78106 78106 0.0 fgets
|
|
0.0 63408 1 63408.0 63408.0 63408 63408 0.0 fflush
|
|
0.0 48070 9 5341.1 4538.0 2655 10159 2577.4 munmap
|
|
0.0 47368 6 7894.7 6457.0 3056 16481 4729.8 open
|
|
0.0 21011 3 7003.7 8336.0 3808 8867 2780.2 pipe2
|
|
0.0 18394 10 1839.4 1417.5 1021 3646 945.4 read
|
|
0.0 16351 2 8175.5 8175.5 6753 9598 2011.7 socket
|
|
0.0 15940 1 15940.0 15940.0 15940 15940 0.0 connect
|
|
0.0 6122 1 6122.0 6122.0 6122 6122 0.0 fopen64
|
|
0.0 3316 2 1658.0 1658.0 1543 1773 162.6 fwrite
|
|
0.0 2034 1 2034.0 2034.0 2034 2034 0.0 bind
|
|
0.0 1142 1 1142.0 1142.0 1142 1142 0.0 fcntl
|
|
|
|
[5/8] Executing 'cuda_api_sum' stats report
|
|
|
|
Time (%) Total Time (ns) Num Calls Avg (ns) Med (ns) Min (ns) Max (ns) StdDev (ns) Name
|
|
-------- --------------- --------- ---------- ---------- -------- --------- ----------- ---------------------------------
|
|
95.9 293359199 3 97786399.7 36170906.0 18784 257169509 139208389.2 cudaMemcpy
|
|
1.7 5152663 12 429388.6 6382.0 190 1823102 617871.7 cudaFree
|
|
0.8 2489648 1 2489648.0 2489648.0 2489648 2489648 0.0 cudaFreeHost
|
|
0.5 1393628 1 1393628.0 1393628.0 1393628 1393628 0.0 cudaGetDeviceProperties_v2_v12000
|
|
0.5 1378710 1 1378710.0 1378710.0 1378710 1378710 0.0 cudaMallocHost
|
|
0.3 970320 171 5674.4 2835.0 2715 346198 26465.2 cudaLaunchKernel
|
|
0.3 874376 6 145729.3 126041.0 3197 338894 137484.8 cudaMalloc
|
|
0.1 155420 838 185.5 150.0 80 4749 179.1 cuGetProcAddress_v2
|
|
0.0 33162 1 33162.0 33162.0 33162 33162 0.0 cudaMemset
|
|
0.0 19137 18 1063.2 301.0 250 8406 1995.0 cudaEventCreateWithFlags
|
|
0.0 13998 4 3499.5 2850.5 1113 7184 2591.7 cudaDeviceSynchronize
|
|
0.0 8342 18 463.4 296.0 210 2375 525.8 cudaEventDestroy
|
|
0.0 5170 4 1292.5 1307.5 1142 1413 132.0 cuInit
|
|
0.0 1463 3 487.7 180.0 110 1173 594.5 cuModuleGetLoadingMode
|
|
0.0 832 2 416.0 416.0 261 571 219.2 cudaGetDriverEntryPoint_v11030
|
|
|
|
[6/8] Executing 'cuda_gpu_kern_sum' stats report
|
|
|
|
Time (%) Total Time (ns) Instances Avg (ns) Med (ns) Min (ns) Max (ns) StdDev (ns) Name
|
|
-------- --------------- --------- --------- --------- -------- -------- ----------- ----------------------------------------------------------------------------------------------------
|
|
97.4 243410880 49 4967569.0 4540231.0 1227010 73760510 10134533.5 matmul_forward_kernel(float *, const float *, const float *, const float *, int, int, int, int)
|
|
1.6 4088678 25 163547.1 163712.0 158528 166049 1938.8 layernorm_forward_kernel(float *, float *, float *, const float *, const float *, const float *, in…
|
|
0.5 1133283 12 94440.3 94448.5 94144 94656 137.9 preattention_scores_kernel(float *, float *, int, int, int, int, float *)
|
|
0.1 345409 12 28784.1 28448.0 28192 29889 625.8 attention_v_kernel(float *, float *, float *, int, int, int, int)
|
|
0.1 227136 12 18928.0 18944.0 18784 19008 57.1 gelu_forward_kernel(float *, const float *, int)
|
|
0.1 215745 12 17978.8 17952.0 17824 18208 112.4 softmax_forward_kernel(float *, float, const float *, int, int)
|
|
0.1 145632 24 6068.0 6080.0 5632 6592 393.7 residual_forward_kernel(float *, float *, float *, int)
|
|
0.0 117472 12 9789.3 9760.0 9632 10080 114.5 permute_kernel(float *, float *, float *, const float *, int, int, int, int)
|
|
0.0 86400 1 86400.0 86400.0 86400 86400 0.0 encoder_forward_kernel(float *, const int *, const float *, const float *, int, int, int)
|
|
0.0 70560 12 5880.0 5888.0 5824 5888 19.9 unpermute_kernel(float *, float *, int, int, int, int)
|
|
|
|
[7/8] Executing 'cuda_gpu_mem_time_sum' stats report
|
|
|
|
Time (%) Total Time (ns) Count Avg (ns) Med (ns) Min (ns) Max (ns) StdDev (ns) Operation
|
|
-------- --------------- ----- ---------- ---------- -------- -------- ----------- ----------------------------
|
|
81.6 36080190 2 18040095.0 18040095.0 864 36079326 25511325.1 [CUDA memcpy Host-to-Device]
|
|
17.5 7741294 1 7741294.0 7741294.0 7741294 7741294 0.0 [CUDA memcpy Device-to-Host]
|
|
0.9 408288 1 408288.0 408288.0 408288 408288 0.0 [CUDA memset]
|
|
|
|
[8/8] Executing 'cuda_gpu_mem_size_sum' stats report
|
|
|
|
Total (MB) Count Avg (MB) Med (MB) Min (MB) Max (MB) StdDev (MB) Operation
|
|
---------- ----- -------- -------- -------- -------- ----------- ----------------------------
|
|
497.905 2 248.952 248.952 0.001 497.904 352.070 [CUDA memcpy Host-to-Device]
|
|
232.392 1 232.392 232.392 232.392 232.392 0.000 [CUDA memset]
|
|
51.511 1 51.511 51.511 51.511 51.511 0.000 [CUDA memcpy Device-to-Host]
|
|
|
|
Generated:
|
|
/u/asingh41/sp26_ece408_Temperature1.5/report1.nsys-rep
|
|
/u/asingh41/sp26_ece408_Temperature1.5/report1.sqlite
|