1. Parallelize grouped low-rank output projection (SumHeads) across NUMA - #1009
Merged
Conversation
copybara-service
Bot
force-pushed
the
test_974965810
branch
from
September 3, 2026 10:11
acb84d4 to
40c6ac1
Compare
clusters via ParallelFor over o_groups. 2. Direct strided matrix view (C_g) into mla_o_mid destination buffer, eliminating intermediate copy per group. 3. Demote att_out directly from float to BF16 into per-cluster mla_o_in buffers, switching mla_o_mid to BF16 to reduce memory footprint and bandwidth. PiperOrigin-RevId: 975630365
copybara-service
Bot
force-pushed
the
test_974965810
branch
from
September 3, 2026 10:20
40c6ac1 to
30b178e
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
clusters via ParallelFor over o_groups.
eliminating intermediate copy per group.
buffers, switching mla_o_mid to BF16 to reduce memory footprint and
bandwidth.