Implement VNNI based flash attention that uses SumOfMulQuadAccumulate#957
Merged
Conversation
copybara-service
Bot
force-pushed
the
test_947705494
branch
7 times, most recently
from
July 22, 2026 22:40
09ab8ae to
cfad1f2
Compare
It uses int8 kv's and on supported target can bring about 2x speedup of attention and about 10% speedup overall during prefill. PiperOrigin-RevId: 952397566
copybara-service
Bot
force-pushed
the
test_947705494
branch
from
July 22, 2026 23:31
cfad1f2 to
f1b8d5e
Compare
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Implement VNNI based flash attention that uses SumOfMulQuadAccumulate
It uses int8 kv's and on supported target can bring about 2x speedup of attention and about 10% speedup overall during prefill.