Refactor 2 awq gemm kernels into m16nXk32 #2723
Merged
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
zcnrex
changed the title
|
@WoosukKwon could you please help review? |
|
@zcnrex Nice job refactoring the GEMM kernels. I am especially excited to see if the dequant + cublas can bring a speedup like we talked about. |
WoosukKwon
approved these changes
Feb 12, 2024
yhu422
added a commit
to yhu422/vllm
that referenced
this pull request
Feb 13, 2024
[ROCm] Fix build problem resulted from previous commit related to FP8 kv-cache support (vllm-project#2790) Add documentation on how to do incremental builds (vllm-project#2796) [Ray] Integration compiled DAG off by default (vllm-project#2471) Disable custom all reduce by default (vllm-project#2808) add usage context removed usage_context from Engine_args Move IO to another process added http request [ROCm] support Radeon™ 7900 series (gfx1100) without using flash-attention (vllm-project#2768) Add documentation section about LoRA (vllm-project#2834) Refactor 2 awq gemm kernels into m16nXk32 (vllm-project#2723) Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net> Added additional arg for from_engine_args comments
jvmncs
pushed a commit
to jvmncs/vllm
that referenced
this pull request
Feb 14, 2024
Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net>
3 tasks
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Feb 20, 2024
Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net>
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Feb 22, 2024
Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net>
xjpang
pushed a commit
to xjpang/vllm
that referenced
this pull request
Mar 4, 2024
Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net>
|
@zcnrex @WoosukKwon |
Temirulan
pushed a commit
to Temirulan/vllm-whisper
that referenced
this pull request
Sep 6, 2024
Co-authored-by: Chunan Zeng <chunanzeng@Chunans-Air.attlocal.net>
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.
Merge
gemm_forward_4bit_cuda_m16n64k32andgemm_forward_4bit_cuda_m16n128k32into 1 function, and pass 64/128 as a parameterUsed the diff tool to facilitate identifying numbers to change:
https://www.diffchecker.com/VIc6ukUz/
Next PRs: add cuBlas as an alternative to awq-gemm as discussed in #2566 (comment)