-
Notifications
You must be signed in to change notification settings - Fork 224
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add a check for packed tensors for convolution solvers #2471
Changes from 2 commits
dfee26c
c3166b9
331b19d
fa00081
a0dcdec
b9734ca
82cf6d7
494b325
d702070
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -44,15 +44,14 @@ bool ConvDirectNaiveConvBwd::IsApplicable(const ExecutionContext& ctx, | |
if(!ConvDirectNaiveConvIsApplicableByKernelType(ctx, problem)) | ||
return false; | ||
|
||
if(!problem.direction.IsBackwardData()) | ||
return false; | ||
Comment on lines
+47
to
+48
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Existing code has been moved There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I will revert it but the existing order of checks itself is something to discuss as it seems arbitrary and based on what the programmer wrote at the time or copy-pasted from somewhere else. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I agree that the current order of checking conditions may not be optimal, but here we need to do a little research and evaluate the expected benefit from this. It's better to do this in a separate pull request |
||
if(!problem.IsLayoutDefault() && !problem.IsLayoutNHWC()) | ||
return false; | ||
|
||
if(!(problem.IsFp32() || problem.IsFp16() || problem.IsBfp16() || problem.IsFp8() || | ||
problem.IsBfp8())) | ||
return false; | ||
|
||
if(!problem.direction.IsBackwardData()) | ||
return false; | ||
if(problem.IsTensorsCasted()) | ||
{ | ||
auto test_cast = [&](const TensorDescriptor& desc) { | ||
|
This comment was marked as resolved.
Sorry, something went wrong.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@atamazov : favoring the current name as it is more descriptive.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
[resolved]