Adding Triton implementations awq_dequantize and awq_gemm to ROCm #136
Conversation
… ransmith_int4_experiments
…t not when otherwise
rasmith
changed the title
rasmith
requested review from
charlifu,
divakar-amd,
gshtras and
hegemanjw4amd
vllm/envs.py
Outdated
|
|
||
| # If set, vllm will print verbose logs during installation | ||
| "VLLM_USE_TRITON_AWQ": | ||
| lambda: bool(int(os.getenv("VLLM_USE_TRITON_AWQ", '0'))), |
There was a problem hiding this comment.
Can we do this as:
lambda: (os.getenv("VLLM_USE_TRITON_AWQ", "True").lower() in ("true", "1"))
(Make the default "True", to favor Triton.)
There was a problem hiding this comment.
Oh yes, that makes sense for our repo.
| split_k_iters, # Not used | ||
| thx, # Not used | ||
| thy, # Not used | ||
| group_size, # Should always be 128 |
There was a problem hiding this comment.
Maybe remove the comment about "always being 128"... This should work for any reasonable value.
There was a problem hiding this comment.
The group size is set when the model is quantized and it's always 128. I can update the comment though.
|
Also, do you want to squash your commits? |
|
@hegemanjw4amd If our repo allows it, I'll just squash-merge, which will just result in one single commit for the whole PR. |
|
Approving; feel free to fix up minor nits as appropriate. |
3 participants
This PR adds basic support for AWQ by adding Triton implementations of awq_dequantize and awq_gemm to ROCm vLLM. Performance is not optimized and is quite slow, instead correctness is prioritized.
BEFORE SUBMITTING, PLEASE READ THE CHECKLIST BELOW AND FILL IN THE DESCRIPTION ABOVE
PR Checklist (Click to Expand)
Thank you for your contribution to vLLM! Before submitting the pull request, please ensure the PR meets the following criteria. This helps vLLM maintain the code quality and improve the efficiency of the review process.
PR Title and Classification
Only specific types of PRs will be reviewed. The PR title is prefixed appropriately to indicate the type of change. Please use one of the following:
[Bugfix]for bug fixes.[CI/Build]for build or continuous integration improvements.[Doc]for documentation fixes and improvements.[Model]for adding a new model or improving an existing model. Model name should appear in the title.[Frontend]For changes on the vLLM frontend (e.g., OpenAI API server,LLMclass, etc.)[Kernel]for changes affecting CUDA kernels or other compute kernels.[Core]for changes in the core vLLM logic (e.g.,LLMEngine,AsyncLLMEngine,Scheduler, etc.)[Hardware][Vendor]for hardware-specific changes. Vendor name should appear in the prefix (e.g.,[Hardware][AMD]).[Misc]for PRs that do not fit the above categories. Please use this sparingly.Note: If the PR spans more than one category, please include all relevant prefixes.
Code Quality
The PR need to meet the following code quality standards:
format.shto format your code.docs/source/if the PR modifies the user-facing behaviors of vLLM. It helps vLLM user understand and utilize the new features or changes.Notes for Large Changes
Please keep the changes as concise as possible. For major architectural changes (>500 LOC excluding kernel/data/config/test), we would expect a GitHub issue (RFC) discussing the technical design and justification. Otherwise, we will tag it with
rfc-requiredand might not go through the PR.What to Expect for the Reviews
The goal of the vLLM team is to be a transparent reviewing machine. We would like to make the review process transparent and efficient and make sure no contributor feel confused or frustrated. However, the vLLM team is small, so we need to prioritize some PRs over others. Here is what you can expect from the review process:
action-requiredlabel on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.Thank You
Finally, thank you for taking the time to read these guidelines and for your interest in contributing to vLLM. Your contributions make vLLM a great tool for everyone!