[Kernel][ROCm][AMD] enable fused topk_softmax kernel for moe layer

[divakar-amd]

This PR enables the fused topk_softmax kernel used in moe layer for HIP

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, LLM class, 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:

• We adhere to Google Python style guide and Google C++ style guide.
• Pass all linter checks. Please use format.sh to format your code.
• The code need to be well-documented to ensure future contributors can easily understand the code.
• Include sufficient tests to ensure the project to stay correct and robust. This includes both unit tests and integration tests.
• Please add documentation to 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-required and 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:

• After the PR is submitted, the PR will be assigned to a reviewer. Every reviewer will pick up the PRs based on their expertise and availability.
• After the PR is assigned, the reviewer will provide status update every 2-3 days. If the PR is not reviewed within 7 days, please feel free to ping the reviewer or the vLLM team.
• After the review, the reviewer will put an action-required label on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.
• Please respond to all comments within a reasonable time frame. If a comment isn't clear or you disagree with a suggestion, feel free to ask for clarification or discuss the suggestion.

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!

[pcmoritz]

This looks like a good change! In order to review this, I'd like to try it out first. I have an account on aac6.amd.com but currently no access to e.g. MI300, if you give me access I'm happy to review this / get it merged :)

[shajrawi]

We don't have external AAC access to MI300 at the moment - we are working with @WoosukKwon , @simon-mo et.al - you guys are on the list. But AAC is not going to be ready soon..

[pcmoritz]

If developer hardware access is not ready yet, adding the code path to the AMD CI could be an option -- this code path is covered in the https://github.com/vllm-project/vllm/blob/main/tests/kernels/test_moe.py test which is part of the kernels tests https://github.com/vllm-project/vllm/blob/main/.buildkite/test-pipeline.yaml#L84. Looks like that's being worked on in #4922 so after that is ready we can just run this PR against the CI :)

[divakar-amd]

Hi @pcmoritz, the PR has passed the CI tests containing AMD test cases. This should be good to merge now.

[pcmoritz]

@divakar-amd I'll merge it, but please note that the "kernels" test is currently not run in the AMD CI -- and the test is in https://github.com/vllm-project/vllm/blob/main/tests/kernels/test_moe.py, so it is not tested in the CI (and I have no way to test it manually).

[divakar-amd]

Thanks @pcmoritz for the merge and sharing the "kernels" test. We will get this included in the AMD CI soon.
Also, I checked the moe test manually for this PR and it passes.