SRFSN: Replace MATMUL intrisics with explicit loops#46
Merged
Conversation
The use of MATMUL intrinsics has proven problematic for GPU compilation with NVHPC-22.11. We've replaced it here with manual SUM and multiplications, but this was not perfectly bit-identical on CPU.
This fixes various CI test runs follwoing the Cy50r3 update.
awnawab
approved these changes
Jun 9, 2026
awnawab
left a comment
awnawab
left a comment
Collaborator
There was a problem hiding this comment.
The changes look good to me ✅ thanks a lot for grinding through what was quite a painful debug session 🙏
Collaborator
|
I think that's ok, I'd say over a multi-year run we can confidently use a lower threshold, which I think it is justified if a relatively small change like this introduces deviations. Thanks a lot @mlange05 , good to go for me. |
gabardu
approved these changes
Jun 9, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
This PR is bringing in one of the more subtle changes in the Cy50r3 merge, which requires adjustments of test error norms for the
insitu_US_Ha1test for certain compilers. It simply replaces a set of MATMUL intrinsics in the snow scheme with equivalent explicit loops and adjusts the error tolerances for specific compilers to get the CI runners to pass.I've also piggy-backed a small fix to an integer KIND declaration in the same file - even though this should be unrelated.
Also tagging @piotrows for visibility.
Contributor Declaration
By opening this pull request, I affirm the following: