Fix Q4_K and Q5_K for QK_K = 64 on CUDA #2359
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
JohannesGaessler
approved these changes
Jul 24, 2023
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Outdated suggestions cannot be applied.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.
They were broken (don't even compile) when
LLAMA_CUDA_FORCE_DMMV=OFF. As per @JohannesGaessler, when added the matrix times vector versions that use integer SIMD intrinsics, he did not implement forQK_K = 64as he consideredQK_K = 64to be a temporary fix. I agree with this sentiment, but in the meantime it is not good to have broken stuff on master, so this PR fixes that.A fix was needed only for
Q4_KandQ5_K.Q2_K,Q3_K, andQ6_Kworked out of the box.Performance is not great, but I did not want to spend the time optimizing a temporary solution.