Fix Vulkan Quantized Matrix Vector Multiplication on AMD GPUs when ncols < 64 #8855
Conversation
This file contains 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
0cc4m
changed the title
0cc4m
changed the title
ggerganov
approved these changes
Aug 5, 2024
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this pull request
Aug 7, 2024
…ov#8855) * Fix Vulkan mul mat vec invalid results when ncols < warp size * Only run backend ops mul mat vec block size test if block size not already covered
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Outdated suggestions cannot be applied.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.
I fixed Vulkan quantized matrix vector multiplication test failure on AMD GPUs (warp size 64) when there are not enough blocks to fill the warp. This was caught by the tests added in #8800 , but I noticed that for k-quants they run the same test twice, so I added a check whether the new test is actually required. Let me know if that's okay.