[Feature] Support AVX2 on linux #192
Open
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
added 3 commits
May 14, 2025 02:01
Signed-off-by: lshAlgorithm <[email protected]>
Signed-off-by: lshAlgorithm <[email protected]>
Signed-off-by: lshAlgorithm <[email protected]>
Signed-off-by: lshAlgorithm <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
Support the AVX2 on Linux.
Main change: In
rwkv_operators_wkv_v7.inc, it is straightforward, we just need to add things up in vector and reduce then to sum out of loop usinghorizontal_sum.Reproduce: I've added the compile option, so just compile it using
./script.sh, and rungenerate_completion.pyfollowing README.Remaining issues: Because the reduction of AVX vector is instrinsics-depedent, and my pc doesn't support avx512, let alone arm, the code support AVX2 ONLY; and it seems that ggml doesn't fully support AVX2, and doesn't include <immintrin.h>. (Maybe I don't get it right, I just add the header in the code.)
Outcome: Test on Fedora40 with Intel(R) Core(TM) Ultra 9 185H (No GPU). Slightly speedup using rwkv-7-1.5B, fp16, so as rwkv-7-2.9B, fp16. (To be honest, it is about ms level faster, and my PC is not that stable, so I cannot show an accurate speedup...)
Statement: Sorry for the stupid commit explanation, it was meant for private use days before. No much experience in PR, but really wanna contribute to the community. Hope for suggestion.