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, -43.363079
|
|
-39.836346, -39.774330
|
|
-43.065910, -43.004066
|
|
-42.828045, -42.763325
|
|
-43.529541, -43.461735
|
|
-44.318398, -44.251251
|
|
-41.227425, -41.161076
|
|
-41.270760, -41.205547
|
|
-42.541393, -42.472733
|
|
-42.394997, -42.328682
|
|
---------------- Logits OK! (max err 0.036241, RMSE 0.001194) ----------------
|
|
_________________________ All tests passed! _________________________
|
|
Collecting data...
|
|
Generating '/tmp/nsys-report-c9b3.qdstrm'
|
|
|
|
[1/8] [0% ] report8.nsys-rep
|
|
[1/8] [0% ] report8.nsys-rep
|
|
[1/8] [====28% ] report8.nsys-rep
|
|
[1/8] [=====29% ] report8.nsys-rep
|
|
[1/8] [=====30% ] report8.nsys-rep
|
|
[1/8] [=======39% ] report8.nsys-rep
|
|
[1/8] [==========47% ] report8.nsys-rep
|
|
[1/8] [==========48% ] report8.nsys-rep
|
|
[1/8] [==========49% ] report8.nsys-rep
|
|
[1/8] [===========50% ] report8.nsys-rep
|
|
[1/8] [===========51% ] report8.nsys-rep
|
|
[1/8] [============57% ] report8.nsys-rep
|
|
[1/8] [=============58% ] report8.nsys-rep
|
|
[1/8] [=============59% ] report8.nsys-rep
|
|
[1/8] [=============60% ] report8.nsys-rep
|
|
[1/8] [==============61% ] report8.nsys-rep
|
|
[1/8] [==============62% ] report8.nsys-rep
|
|
[1/8] [==============63% ] report8.nsys-rep
|
|
[1/8] [=================74% ] report8.nsys-rep
|
|
[1/8] [=======================94% ] report8.nsys-rep
|
|
[1/8] [========================100%] report8.nsys-rep
|
|
[1/8] [========================100%] report8.nsys-rep
|
|
|
|
[2/8] [0% ] report8.sqlite
|
|
[2/8] [1% ] report8.sqlite
|
|
[2/8] [2% ] report8.sqlite
|
|
[2/8] [3% ] report8.sqlite
|
|
[2/8] [4% ] report8.sqlite
|
|
[2/8] [5% ] report8.sqlite
|
|
[2/8] [6% ] report8.sqlite
|
|
[2/8] [7% ] report8.sqlite
|
|
[2/8] [8% ] report8.sqlite
|
|
[2/8] [9% ] report8.sqlite
|
|
[2/8] [10% ] report8.sqlite
|
|
[2/8] [11% ] report8.sqlite
|
|
[2/8] [12% ] report8.sqlite
|
|
[2/8] [13% ] report8.sqlite
|
|
[2/8] [14% ] report8.sqlite
|
|
[2/8] [=15% ] report8.sqlite
|
|
[2/8] [=16% ] report8.sqlite
|
|
[2/8] [=17% ] report8.sqlite
|
|
[2/8] [==18% ] report8.sqlite
|
|
[2/8] [==19% ] report8.sqlite
|
|
[2/8] [==20% ] report8.sqlite
|
|
[2/8] [==21% ] report8.sqlite
|
|
[2/8] [===22% ] report8.sqlite
|
|
[2/8] [===23% ] report8.sqlite
|
|
[2/8] [===24% ] report8.sqlite
|
|
[2/8] [====25% ] report8.sqlite
|
|
[2/8] [====26% ] report8.sqlite
|
|
[2/8] [====27% ] report8.sqlite
|
|
[2/8] [====28% ] report8.sqlite
|
|
[2/8] [=====29% ] report8.sqlite
|
|
[2/8] [=====30% ] report8.sqlite
|
|
[2/8] [=====31% ] report8.sqlite
|
|
[2/8] [=====32% ] report8.sqlite
|
|
[2/8] [======33% ] report8.sqlite
|
|
[2/8] [======34% ] report8.sqlite
|
|
[2/8] [======35% ] report8.sqlite
|
|
[2/8] [=======36% ] report8.sqlite
|
|
[2/8] [=======37% ] report8.sqlite
|
|
[2/8] [=======38% ] report8.sqlite
|
|
[2/8] [=======39% ] report8.sqlite
|
|
[2/8] [========40% ] report8.sqlite
|
|
[2/8] [========41% ] report8.sqlite
|
|
[2/8] [========42% ] report8.sqlite
|
|
[2/8] [=========43% ] report8.sqlite
|
|
[2/8] [=========44% ] report8.sqlite
|
|
[2/8] [=========45% ] report8.sqlite
|
|
[2/8] [=========46% ] report8.sqlite
|
|
[2/8] [==========47% ] report8.sqlite
|
|
[2/8] [==========48% ] report8.sqlite
|
|
[2/8] [==========49% ] report8.sqlite
|
|
[2/8] [===========50% ] report8.sqlite
|
|
[2/8] [===========51% ] report8.sqlite
|
|
[2/8] [===========52% ] report8.sqlite
|
|
[2/8] [===========53% ] report8.sqlite
|
|
[2/8] [============54% ] report8.sqlite
|
|
[2/8] [============55% ] report8.sqlite
|
|
[2/8] [============56% ] report8.sqlite
|
|
[2/8] [============57% ] report8.sqlite
|
|
[2/8] [=============58% ] report8.sqlite
|
|
[2/8] [=============59% ] report8.sqlite
|
|
[2/8] [=============60% ] report8.sqlite
|
|
[2/8] [==============61% ] report8.sqlite
|
|
[2/8] [==============62% ] report8.sqlite
|
|
[2/8] [==============63% ] report8.sqlite
|
|
[2/8] [==============64% ] report8.sqlite
|
|
[2/8] [===============65% ] report8.sqlite
|
|
[2/8] [===============66% ] report8.sqlite
|
|
[2/8] [===============67% ] report8.sqlite
|
|
[2/8] [================68% ] report8.sqlite
|
|
[2/8] [================69% ] report8.sqlite
|
|
[2/8] [================70% ] report8.sqlite
|
|
[2/8] [================71% ] report8.sqlite
|
|
[2/8] [=================72% ] report8.sqlite
|
|
[2/8] [=================73% ] report8.sqlite
|
|
[2/8] [=================74% ] report8.sqlite
|
|
[2/8] [==================75% ] report8.sqlite
|
|
[2/8] [==================76% ] report8.sqlite
|
|
[2/8] [==================77% ] report8.sqlite
|
|
[2/8] [==================78% ] report8.sqlite
|
|
[2/8] [===================79% ] report8.sqlite
|
|
[2/8] [===================80% ] report8.sqlite
|
|
[2/8] [===================81% ] report8.sqlite
|
|
[2/8] [===================82% ] report8.sqlite
|
|
[2/8] [====================83% ] report8.sqlite
|
|
[2/8] [====================84% ] report8.sqlite
|
|
[2/8] [====================85% ] report8.sqlite
|
|
[2/8] [=====================86% ] report8.sqlite
|
|
[2/8] [=====================87% ] report8.sqlite
|
|
[2/8] [=====================88% ] report8.sqlite
|
|
[2/8] [=====================89% ] report8.sqlite
|
|
[2/8] [======================90% ] report8.sqlite
|
|
[2/8] [======================91% ] report8.sqlite
|
|
[2/8] [======================92% ] report8.sqlite
|
|
[2/8] [=======================93% ] report8.sqlite
|
|
[2/8] [=======================94% ] report8.sqlite
|
|
[2/8] [=======================95% ] report8.sqlite
|
|
[2/8] [=======================96% ] report8.sqlite
|
|
[2/8] [========================97% ] report8.sqlite
|
|
[2/8] [========================98% ] report8.sqlite
|
|
[2/8] [========================99% ] report8.sqlite
|
|
[2/8] [========================100%] report8.sqlite
|
|
[2/8] [========================100%] report8.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
|
|
-------- --------------- --------- ----------- ----------- -------- --------- ----------- ---------------------
|
|
54.7 1713974843 27 63480549.7 100113755.0 1193 108696839 46506883.0 poll
|
|
36.4 1138051148 6 189675191.3 15085903.5 7013 820460548 328760934.6 fread
|
|
4.9 153194333 595 257469.5 22662.0 1032 18701261 996184.4 ioctl
|
|
3.8 118955136 38 3130398.3 6247.0 3576 118016722 19140872.4 fopen
|
|
0.1 2406150 25 96246.0 8916.0 6703 1662941 329260.7 mmap64
|
|
0.1 1648944 10 164894.4 30456.5 21350 809013 285878.4 write
|
|
0.0 781242 32 24413.8 2209.0 1333 381844 85901.8 fclose
|
|
0.0 578514 17 34030.2 4980.0 2104 377536 90206.1 mmap
|
|
0.0 413214 2 206607.0 206607.0 163236 249978 61335.9 pthread_create
|
|
0.0 322402 55 5861.9 5991.0 2354 11190 2184.1 open64
|
|
0.0 76614 1 76614.0 76614.0 76614 76614 0.0 fgets
|
|
0.0 61595 1 61595.0 61595.0 61595 61595 0.0 fflush
|
|
0.0 43209 8 5401.1 4368.0 2996 11862 2985.1 munmap
|
|
0.0 38103 6 6350.5 6252.0 3076 10069 2427.2 open
|
|
0.0 24296 3 8098.7 7855.0 3567 12874 4658.3 pipe2
|
|
0.0 18094 9 2010.4 1663.0 1032 3627 895.0 read
|
|
0.0 16701 2 8350.5 8350.5 5090 11611 4611.0 socket
|
|
0.0 11391 1 11391.0 11391.0 11391 11391 0.0 connect
|
|
0.0 6843 1 6843.0 6843.0 6843 6843 0.0 fopen64
|
|
0.0 3707 2 1853.5 1853.5 1453 2254 566.4 fwrite
|
|
0.0 2465 1 2465.0 2465.0 2465 2465 0.0 bind
|
|
0.0 2114 2 1057.0 1057.0 1042 1072 21.2 fcntl
|
|
0.0 1062 1 1062.0 1062.0 1062 1062 0.0 listen
|
|
0.0 1042 1 1042.0 1042.0 1042 1042 0.0 pthread_mutex_trylock
|
|
|
|
[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
|
|
-------- --------------- --------- ---------- ---------- -------- -------- ----------- ---------------------------------
|
|
75.3 55764098 3 18588032.7 18908389.0 17473 36838236 18412471.8 cudaMemcpy
|
|
13.8 10194027 12 849502.3 11000.0 180 5616656 1649151.5 cudaFree
|
|
3.3 2407413 1 2407413.0 2407413.0 2407413 2407413 0.0 cudaFreeHost
|
|
2.2 1633015 1 1633015.0 1633015.0 1633015 1633015 0.0 cudaGetDeviceProperties_v2_v12000
|
|
1.8 1328725 1 1328725.0 1328725.0 1328725 1328725 0.0 cudaMallocHost
|
|
1.5 1095220 171 6404.8 3026.0 2846 406250 31186.8 cudaLaunchKernel
|
|
1.1 837937 6 139656.2 131245.5 4819 364852 133988.9 cudaMalloc
|
|
0.6 426397 18 23688.7 295.0 260 416068 97927.8 cudaEventCreateWithFlags
|
|
0.3 225714 838 269.3 180.0 90 44584 1538.5 cuGetProcAddress_v2
|
|
0.1 59281 1 59281.0 59281.0 59281 59281 0.0 cudaMemset
|
|
0.0 21260 4 5315.0 3978.0 1162 12142 4779.4 cudaDeviceSynchronize
|
|
0.0 9335 18 518.6 290.5 210 3466 760.8 cudaEventDestroy
|
|
0.0 5982 4 1495.5 1513.0 1293 1663 172.9 cuInit
|
|
0.0 1283 2 641.5 641.5 261 1022 538.1 cudaGetDriverEntryPoint_v11030
|
|
0.0 1233 3 411.0 131.0 90 1012 520.9 cuModuleGetLoadingMode
|
|
|
|
[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
|
|
-------- --------------- --------- -------- -------- -------- -------- ----------- ----------------------------------------------------------------------------------------------------
|
|
90.7 10611176 49 216554.6 162239.0 77280 2391507 328408.1 matmul_forward_kernel(float *, const float *, const float *, const float *, int, int, int, int)
|
|
1.9 226655 12 18887.9 18880.0 18784 18944 41.2 gelu_forward_kernel(float *, const float *, int)
|
|
1.2 145759 24 6073.3 6080.0 5663 6528 378.6 residual_forward_kernel(float *, float *, float *, int)
|
|
1.2 141087 12 11757.3 11728.0 11648 12000 97.8 softmax_forward_kernel(float *, float, const float *, int, int)
|
|
1.1 131199 25 5248.0 5088.0 4672 5984 440.1 layernorm_forward_kernel(float *, float *, float *, const float *, const float *, const float *, in…
|
|
0.9 99744 12 8312.0 8288.0 8224 8448 62.7 attention_v_kernel(float *, float *, float *, int, int, int, int)
|
|
0.8 95968 12 7997.3 7968.0 7904 8160 72.8 preattention_scores_kernel(float *, float *, int, int, int, int, float *)
|
|
0.8 94622 12 7885.2 7856.0 7648 8255 167.1 permute_kernel(float *, float *, float *, const float *, int, int, int, int)
|
|
0.7 86176 1 86176.0 86176.0 86176 86176 0.0 encoder_forward_kernel(float *, const int *, const float *, const float *, int, int, int)
|
|
0.6 70623 12 5885.3 5888.0 5856 5920 21.2 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
|
|
-------- --------------- ----- ---------- ---------- -------- -------- ----------- ----------------------------
|
|
81.9 36690814 2 18345407.0 18345407.0 864 36689950 25943101.5 [CUDA memcpy Host-to-Device]
|
|
17.2 7698135 1 7698135.0 7698135.0 7698135 7698135 0.0 [CUDA memcpy Device-to-Host]
|
|
0.9 407997 1 407997.0 407997.0 407997 407997 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/report8.nsys-rep
|
|
/u/asingh41/sp26_ece408_Temperature1.5/report8.sqlite
|