Skip to content

Instantly share code, notes, and snippets.

@colesbury
Created May 8, 2018 17:15
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save colesbury/ea08f4300a52133dfd5201351fcf3b07 to your computer and use it in GitHub Desktop.
Save colesbury/ea08f4300a52133dfd5201351fcf3b07 to your computer and use it in GitHub Desktop.
------------------------------------- --------------- --------------- --------------- --------------- ---------------
Name CPU time CUDA time Calls CPU total CUDA total
------------------------------------- --------------- --------------- --------------- --------------- ---------------
N5torch8autograd9GraphRootE 52.397us 1.408us 1 52.397us 1.408us
expand 6.042us 1.584us 2 12.084us 3.168us
fill_ 10.903us 4.384us 1 10.903us 4.384us
TBackward 22.800us 4.224us 2 45.599us 8.448us
view 7.472us 4.352us 2 14.945us 8.704us
log_softmax_backward_data 16.972us 8.800us 1 16.972us 8.800us
ones_like 23.823us 10.048us 1 23.823us 10.048us
log_softmax 15.925us 11.424us 1 15.925us 11.424us
LogSoftmaxBackward 24.864us 11.616us 1 24.864us 11.616us
ViewBackward 12.660us 12.544us 1 12.660us 12.544us
nll_loss_forward 16.324us 13.664us 1 16.324us 13.664us
_sum 19.771us 7.104us 2 39.542us 14.208us
nll_loss 24.184us 16.768us 1 24.184us 16.768us
sum 26.483us 9.904us 2 52.967us 19.808us
t 4.586us 1.990us 10 45.857us 19.904us
nll_loss_backward 33.871us 24.992us 1 33.871us 24.992us
ExpandBackward 31.878us 12.736us 2 63.757us 25.472us
NllLossBackward 49.489us 27.808us 1 49.489us 27.808us
addmm 24.109us 24.768us 2 48.218us 49.536us
set_ 11.587us 28.080us 2 23.174us 56.160us
tensor 8.429us 8.463us 11 92.716us 93.088us
mul_ 8.463us 4.731us 23 194.649us 108.807us
_mm 16.882us 27.584us 4 67.529us 110.336us
N5torch8autograd14AccumulateGradE 8.188us 5.369us 23 188.315us 123.495us
mm 23.375us 31.304us 4 93.499us 125.216us
AddmmBackward 84.599us 81.488us 2 169.199us 162.976us
max_pool2d_forward 18.003us 69.429us 3 54.010us 208.288us
max_pool2d 24.010us 72.352us 3 72.031us 217.056us
pin_memory 842.488us 111.520us 2 1684.976us 223.040us
add_ 7.712us 4.425us 69 532.153us 305.319us
prelu_forward 18.348us 58.254us 7 128.438us 407.776us
prelu 24.549us 61.088us 7 171.844us 427.616us
max_pool2d_backward 14.878us 160.309us 3 44.634us 480.928us
MaxPool2DBackward 22.470us 164.021us 3 67.410us 492.064us
prelu_backward 1965.621us 346.583us 7 13759.346us 2426.081us
PreluBackward 1976.626us 351.465us 7 13836.380us 2460.257us
cudnn_convolution 58.614us 660.725us 6 351.683us 3964.352us
_convolution 72.559us 667.643us 6 435.356us 4005.856us
convolution 76.873us 670.715us 6 461.236us 4024.288us
conv2d 81.385us 673.781us 6 488.311us 4042.688us
cudnn_convolution_backward 89.318us 1573.344us 6 535.910us 9440.064us
CudnnConvolutionBackward 106.752us 1577.109us 6 640.510us 9462.656us
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment