BUG: Fix float16 einsum fastpaths using wrong tempvar #20462
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
This hopefully tests things well enough, at least some/most of the paths get triggered and led to errors without the previous float16 typing fixes. I manually confirmed that all paths that were *modified* in the previous commit actually get hit with float16 specialized loops. NOTE: This test may be a bit fragile with floating point roundoff errors, and can in parts be relaxed if this happens.
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Outdated suggestions cannot be applied.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.
Backport of #20310.
Fix for gh-20305 @Qiyu8 maybe you know where to best add some tests? I was hoping to add some parametrization and be done with it, but didn't really find the right test to parametrize for float16 that would fail without the fixes.
EDIT: In case anyone ever stumbles on this. The previous tests probably did run all the code paths, the problem is the missing fractional part. Using the wrong tempvar here, meant the accumulation was done in (u)int16. Without a fractional part, that ends up with the correct result though.