109 lines
16 KiB
Plaintext
109 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, -43.379311
|
|
-39.836346, -39.790001
|
|
-43.065910, -43.020863
|
|
-42.828045, -42.780563
|
|
-43.529541, -43.479824
|
|
-44.318398, -44.268772
|
|
-41.227425, -41.177418
|
|
-41.270760, -41.221939
|
|
-42.541393, -42.488613
|
|
-42.394997, -42.345592
|
|
---------------- Logits OK! (max err 0.055688, RMSE 0.001083) ----------------
|
|
_________________________ All tests passed! _________________________
|
|
Collecting data...
|
|
Generating '/tmp/nsys-report-c915.qdstrm'
|
|
|
|
[1/8] [0% ] report6.nsys-rep
|
|
[1/8] [0% ] report6.nsys-rep
|
|
[1/8] [==19% ] report6.nsys-rep
|
|
[1/8] [==21% ] report6.nsys-rep
|
|
[1/8] [===22% ] report6.nsys-rep
|
|
[1/8] [======35% ] report6.nsys-rep
|
|
[1/8] [=======36% ] report6.nsys-rep
|
|
[1/8] [==========49% ] report6.nsys-rep
|
|
[1/8] [===========50% ] report6.nsys-rep
|
|
[1/8] [===========52% ] report6.nsys-rep
|
|
[1/8] [===========53% ] report6.nsys-rep
|
|
[1/8] [============54% ] report6.nsys-rep
|
|
[1/8] [============55% ] report6.nsys-rep
|
|
[1/8] [============56% ] report6.nsys-rep
|
|
[1/8] [============57% ] report6.nsys-rep
|
|
[1/8] [=============58% ] report6.nsys-rep
|
|
[1/8] [=============59% ] report6.nsys-rep
|
|
[1/8] [==================75% ] report6.nsys-rep
|
|
[1/8] [======================91% ] report6.nsys-rep
|
|
[1/8] [========================100%] report6.nsys-rep
|
|
[1/8] [========================100%] report6.nsys-rep
|
|
|
|
[2/8] [0% ] report6.sqlite
|
|
[2/8] [1% ] report6.sqlite
|
|
[2/8] [2% ] report6.sqlite
|
|
[2/8] [3% ] report6.sqlite
|
|
[2/8] [4% ] report6.sqlite
|
|
[2/8] [5% ] report6.sqlite
|
|
[2/8] [6% ] report6.sqlite
|
|
[2/8] [7% ] report6.sqlite
|
|
[2/8] [8% ] report6.sqlite
|
|
[2/8] [9% ] report6.sqlite
|
|
[2/8] [10% ] report6.sqlite
|
|
[2/8] [11% ] report6.sqlite
|
|
[2/8] [12% ] report6.sqlite
|
|
[2/8] [13% ] report6.sqlite
|
|
[2/8] [14% ] report6.sqlite
|
|
[2/8] [=15% ] report6.sqlite
|
|
[2/8] [=16% ] report6.sqlite
|
|
[2/8] [=17% ] report6.sqlite
|
|
[2/8] [==18% ] report6.sqlite
|
|
[2/8] [==19% ] report6.sqlite
|
|
[2/8] [==20% ] report6.sqlite
|
|
[2/8] [==21% ] report6.sqlite
|
|
[2/8] [===22% ] report6.sqlite
|
|
[2/8] [===23% ] report6.sqlite
|
|
[2/8] [===24% ] report6.sqlite
|
|
[2/8] [====25% ] report6.sqlite
|
|
[2/8] [====26% ] report6.sqlite
|
|
[2/8] [====27% ] report6.sqlite
|
|
[2/8] [====28% ] report6.sqlite
|
|
[2/8] [=====29% ] report6.sqlite
|
|
[2/8] [=====30% ] report6.sqlite
|
|
[2/8] [=====31% ] report6.sqlite
|
|
[2/8] [=====32% ] report6.sqlite
|
|
[2/8] [======33% ] report6.sqlite
|
|
[2/8] [======34% ] report6.sqlite
|
|
[2/8] [======35% ] report6.sqlite
|
|
[2/8] [=======36% ] report6.sqlite
|
|
[2/8] [=======37% ] report6.sqlite
|
|
[2/8] [=======38% ] report6.sqlite
|
|
[2/8] [=======39% ] report6.sqlite
|
|
[2/8] [========40% ] report6.sqlite
|
|
[2/8] [========41% ] report6.sqlite
|
|
[2/8] [========42% ] report6.sqlite
|
|
[2/8] [=========43% ] report6.sqlite
|
|
[2/8] [=========44% ] report6.sqlite
|
|
[2/8] [=========45% ] report6.sqlite
|
|
[2/8] [=========46% ] report6.sqlite
|
|
[2/8] [==========47% ] report6.sqlite
|
|
[2/8] [==========48% ] report6.sqlite
|
|
[2/8] [==========49% ] report6.sqlite
|
|
[2/8] [===========50% ] report6.sqlite
|
|
[2/8] [===========51% ] report6.sqlite
|
|
[2/8] [===========52% ] report6.sqlite
|
|
[2/8] [===========53% ] report6.sqlite
|
|
[2/8] [============54% ] report6.sqlite
|
|
[2/8] [============55% ] report6.sqlite
|
|
[2/8] [============56% ] report6.sqlite
|
|
[2/8] [============57% ] report6.sqlite
|
|
[2/8] [=============58% ] report6.sqlite
|
|
[2/8] [=============59% ] report6.sqlite
|
|
[2/8] [=============60% ] report6.sqlite
|
|
[2/8] [==============61% ] report6.sqlite
|
|
[2/8] [==============62% ] report6.sqlite
|
|
[2/8] [==============63% ] report6.sqlite
|
|
[2/8] [==============64% ] report6.sqlite
|
|
[2/8] [===============65% ] report6.sqlite
|
|
[2/8] [===============66% ] report6.sqlite
|
|
[2/8] [===============67% ] report6.sqlite
|
|
[2/8] [================68% ] report6.sqlite
|
|
[2/8] [================69% ] report6.sqlite
|
|
[2/8] [================70% ] report6.sqlite
|
|
[2/8] [================71% ] report6.sqlite
|
|
[2/8] [=================72% ] report6.sqlite
|
|
[2/8] [=================73% ] report6.sqlite
|
|
[2/8] [=================74% ] report6.sqlite
|
|
[2/8] [==================75% ] report6.sqlite
|
|
[2/8] [==================76% ] report6.sqlite
|
|
[2/8] [==================77% ] report6.sqlite
|
|
[2/8] [==================78% ] report6.sqlite
|
|
[2/8] [===================79% ] report6.sqlite
|
|
[2/8] [===================80% ] report6.sqlite
|
|
[2/8] [===================81% ] report6.sqlite
|
|
[2/8] [===================82% ] report6.sqlite
|
|
[2/8] [====================83% ] report6.sqlite
|
|
[2/8] [====================84% ] report6.sqlite
|
|
[2/8] [====================85% ] report6.sqlite
|
|
[2/8] [=====================86% ] report6.sqlite
|
|
[2/8] [=====================87% ] report6.sqlite
|
|
[2/8] [=====================88% ] report6.sqlite
|
|
[2/8] [=====================89% ] report6.sqlite
|
|
[2/8] [======================90% ] report6.sqlite
|
|
[2/8] [======================91% ] report6.sqlite
|
|
[2/8] [======================92% ] report6.sqlite
|
|
[2/8] [=======================93% ] report6.sqlite
|
|
[2/8] [=======================94% ] report6.sqlite
|
|
[2/8] [=======================95% ] report6.sqlite
|
|
[2/8] [=======================96% ] report6.sqlite
|
|
[2/8] [========================97% ] report6.sqlite
|
|
[2/8] [========================98% ] report6.sqlite
|
|
[2/8] [========================99% ] report6.sqlite
|
|
[2/8] [========================100%] report6.sqlite
|
|
[2/8] [========================100%] report6.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
|
|
-------- --------------- --------- ----------- ----------- -------- ---------- ----------- ----------------------
|
|
53.4 2194351837 32 68573494.9 100124295.0 1886 100179994 44439732.4 poll
|
|
42.9 1763348095 7 251906870.7 58081562.0 1467 1200215470 444651839.0 fread
|
|
2.9 119818701 600 199697.8 24235.0 1886 22988850 990232.3 ioctl
|
|
0.5 21655089 38 569870.8 7543.0 4819 20367773 3300181.2 fopen
|
|
0.1 2467848 25 98713.9 10057.0 7124 1710695 338493.4 mmap64
|
|
0.0 1330620 11 120965.5 45816.0 2375 925189 267079.5 write
|
|
0.0 871902 32 27246.9 3736.5 2374 383499 92136.2 fclose
|
|
0.0 634441 17 37320.1 9009.0 3353 404242 96596.2 mmap
|
|
0.0 389366 2 194683.0 194683.0 194648 194718 49.5 pthread_create
|
|
0.0 386078 55 7019.6 7193.0 3422 11105 1929.7 open64
|
|
0.0 135356 32 4229.9 1397.0 1396 91283 15885.4 fgets
|
|
0.0 107831 56 1925.6 1885.0 1397 12781 1497.9 fcntl
|
|
0.0 63625 2 31812.5 31812.5 1396 62229 43015.4 fflush
|
|
0.0 62369 9 6929.9 6635.0 4330 10127 1985.0 munmap
|
|
0.0 49378 6 8229.7 8171.5 4680 12502 2769.2 open
|
|
0.0 49098 34 1444.1 1431.5 1396 1886 85.5 pthread_mutex_trylock
|
|
0.0 41764 15 2784.3 2374.0 1886 5169 1140.3 read
|
|
0.0 23746 3 7915.3 8940.0 4819 9987 2732.1 pipe2
|
|
0.0 21929 2 10964.5 10964.5 8660 13269 3259.1 socket
|
|
0.0 11105 1 11105.0 11105.0 11105 11105 0.0 connect
|
|
0.0 10965 7 1566.4 1467.0 1396 1886 220.6 dup
|
|
0.0 7613 1 7613.0 7613.0 7613 7613 0.0 fopen64
|
|
0.0 6216 2 3108.0 3108.0 2933 3283 247.5 fwrite
|
|
0.0 4190 3 1396.7 1397.0 1396 1397 0.6 pthread_cond_broadcast
|
|
0.0 3841 1 3841.0 3841.0 3841 3841 0.0 bind
|
|
0.0 1885 1 1885.0 1885.0 1885 1885 0.0 listen
|
|
|
|
[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
|
|
-------- --------------- --------- ---------- ---------- -------- -------- ----------- ---------------------------------
|
|
73.6 59403870 3 19801290.0 22548780.0 21232 36833858 18559468.9 cudaMemcpy
|
|
14.8 11937011 12 994750.9 11280.0 1397 7337536 2110925.3 cudaFree
|
|
3.0 2437533 1 2437533.0 2437533.0 2437533 2437533 0.0 cudaFreeHost
|
|
2.4 1975324 1 1975324.0 1975324.0 1975324 1975324 0.0 cudaGetDeviceProperties_v2_v12000
|
|
1.7 1380415 1 1380415.0 1380415.0 1380415 1380415 0.0 cudaMallocHost
|
|
1.6 1294437 171 7569.8 4261.0 3702 400959 30590.6 cudaLaunchKernel
|
|
1.4 1159422 838 1383.6 1397.0 489 50775 1808.0 cuGetProcAddress_v2
|
|
1.2 974146 6 162357.7 144885.5 4749 353537 148396.1 cudaMalloc
|
|
0.1 52451 1 52451.0 52451.0 52451 52451 0.0 cudaMemset
|
|
0.0 36526 18 2029.2 1397.0 1327 7682 1573.8 cudaEventCreateWithFlags
|
|
0.0 34502 4 8625.5 7228.5 4400 15645 4884.7 cudaDeviceSynchronize
|
|
0.0 29613 18 1645.2 1466.0 978 4749 815.2 cudaEventDestroy
|
|
0.0 9919 4 2479.8 2375.0 2305 2864 258.3 cuInit
|
|
0.0 5238 3 1746.0 1397.0 978 2863 989.8 cuModuleGetLoadingMode
|
|
0.0 3283 2 1641.5 1641.5 1397 1886 345.8 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
|
|
-------- --------------- --------- -------- -------- -------- -------- ----------- ----------------------------------------------------------------------------------------------------
|
|
67.6 10665692 49 217667.2 162496.0 77632 2419814 332250.3 matmul_forward_kernel(float *, const float *, const float *, const float *, int, int, int, int)
|
|
25.9 4086409 25 163456.4 163681.0 158656 166305 1957.5 layernorm_forward_kernel(float *, float *, float *, const float *, const float *, const float *, in…
|
|
1.4 226592 12 18882.7 18880.0 18720 19008 64.7 gelu_forward_kernel(float *, const float *, int)
|
|
1.4 215201 12 17933.4 17920.0 17856 18112 72.7 softmax_forward_kernel(float *, float, const float *, int, int)
|
|
0.9 145057 24 6044.0 6064.0 5664 6560 372.4 residual_forward_kernel(float *, float *, float *, int)
|
|
0.6 99136 12 8261.3 8224.0 8160 8512 105.5 attention_v_kernel(float *, float *, float *, int, int, int, int)
|
|
0.6 95616 12 7968.0 7968.0 7904 8032 36.1 preattention_scores_kernel(float *, float *, int, int, int, int, float *)
|
|
0.6 94912 12 7909.3 7888.0 7776 8128 121.2 permute_kernel(float *, float *, float *, const float *, int, int, int, int)
|
|
0.6 87072 1 87072.0 87072.0 87072 87072 0.0 encoder_forward_kernel(float *, const int *, const float *, const float *, int, int, int)
|
|
0.4 70528 12 5877.3 5872.0 5856 5920 24.9 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
|
|
-------- --------------- ----- ---------- ---------- -------- -------- ----------- ----------------------------
|
|
81.9 36688662 2 18344331.0 18344331.0 896 36687766 25941534.6 [CUDA memcpy Host-to-Device]
|
|
17.2 7687347 1 7687347.0 7687347.0 7687347 7687347 0.0 [CUDA memcpy Device-to-Host]
|
|
0.9 407265 1 407265.0 407265.0 407265 407265 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/report6.nsys-rep
|
|
/u/asingh41/sp26_ece408_Temperature1.5/report6.sqlite
|