Repository navigation
[CUDA] Fallback QMM - #3315
Merged
Merged
[CUDA] Fallback QMM#3315
Conversation
1 task done
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This is an implementation of QMM that is supposed to work for all devices and options, it is slow and mainly serves a fallback for cases that we are not optimizing for.
On A100 it has only about half memory bandwidth of the pipelined version.
FP16xINT4 with K-major weights:
FP16xINT4 with N-major weights:
This PR does not implement 3/5/6-bit quants and arbitrary K size, which I'll do in followup PRs.
Also it takes a lot of RAM to build the kernels, and I have switched to large arm runner with swap enabled for running tests.