CUDA: fix matrix multiplication logic for tests #6667
Merged
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
slaren
approved these changes
Apr 13, 2024
tybalex
pushed a commit
to rubra-ai/tools.cpp
that referenced
this pull request
Apr 17, 2024
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Outdated suggestions cannot be applied.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.
This PR fixes
tests/test-backend-opson Pascal. The issue is that the tests pass an instance ofsrc1withGGML_TYPE_F16. This does not happen during actual inference and was therefore not considered for the logic for selecting a matrix multiplication kernel. For Pascal cards other than the P100 there is no fast evaluation code available either. But since this is a case that does not appear during actual inference it's fine to use the slow code for the tests only. Ceveat: (I think) the tests still fail on Maxwell but honestly I don't think it would be worth the effort to investigate and fix.