CUDA: skip compilation of superfluous FA kernels - #21768
Merged
Merged
Conversation
github-actions
Bot
added
Nvidia GPU
am17an
approved these changes
Apr 11, 2026
IMbackK
approved these changes
Apr 11, 2026
oussamaahmia
pushed a commit
to oussamaahmia/llama-cpp-turboquant-gemma4
that referenced
this pull request
Apr 13, 2026
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
This was referenced Jun 11, 2026
fukuro-kun
pushed a commit
to fukuro-kun/fukuro-llama-cpp-turboquant
that referenced
this pull request
Jul 5, 2026
MrLordCat
referenced
this pull request
in MrLordCat/llama.cpp-rdna-lab
Jul 16, 2026
zommiommy
pushed a commit
to zommiommy/llama.cpp
that referenced
this pull request
Aug 18, 2026
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixup to #20998 .
The compilation of FA kernels with head size 512 is supposed to be skipped for GQA ratios of 1 and 2 because those are never used. However, because the invocation of the corresponding template specializations is not guarded with an
if constexprthey are being compiled regardless; this PR adds them. On my server with a 64 core EPYC CPU the total compilation time of the full project without CCache goes down from 330s to 300s.Requirements