CUDA: fix negative KV_max values in FA #15321
Merged
Conversation
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
slaren
approved these changes
Aug 14, 2025
Collaborator
Author
|
The conditions for triggering the issue seem to be inconsistent and I misinterpreted the results I got. |
github-actions
bot
added
Nvidia GPU
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
Looking at the code I noticed that, depending on the inputs, there are scenarios where the values in
KV_maxcould become negative. For this to happen there would need to be 16 consecutive tokens that are completely masked out, and the physical batch size would need to be >= 1024. I'm not sure this is going to fix #15294 or #15112 but it should be fixed either way since it could be causing problems in the future.