Correctly use dtype when computing shared memory requirements of separable convolution #409
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
calling code was accidentally passing image_dtype to the unused anchor argument which caused the default cp.float32 dtype to always be used in the shared memory calculation. This could result in ptxas errors about too much shared data when the itemsize of the dtype was > than for float32.
…eption occured The ResourceLimitError can be checked very quickly, so we want to keep it as well.
grlee77
added
bug
Only warn on shared memory limit if algorithm='shared_memory' was explicitly requested.
|
rerun tests |
jakirkham
approved these changes
Oct 4, 2022
|
Thanks Greg! 🙏 Happy with merging. Only question would be if we can include a test for the error. Could happen in this PR or after |
gigony
approved these changes
Oct 4, 2022
prior to the fix, test failures occured on my system for float64 here for the sigma values 16, 21, 26, 31
Test case has now been added (values aren't checked, it is just to verify that no compilation error occurs). I verified that without the fix in this MR, the dtype=float64 case fails for the parameterized sigmas 16, 21, 26 and 31 on my system. |
jakirkham
approved these changes
Oct 5, 2022
|
@gpucibot merge |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
closes #408
An error in the keyword arguments passed to the shared memory calculation is fixed in the first commit here. In the second commit, a second fallback is added in case of any other anticipated compilation failure.