[Kernel] Refactor CUTLASS kernels to always take scales that reside on the GPU

[tlrmchlsmth]

This PR smooths out the way the kernels handle per tensor vs per token (or per output channel) scales. Previously, the CUTLASS epilogues needed to be passed a float* for the per-token case and a float for the per-tensor case. The code that called the CUTLASS GEMMs handled this by taking torch::tensors as arguments and unpacking them, calling tensor.item on the scales for the per-tensor case. This was problematic for a few reasons:

• This caused a performance hazard where the scales needed to be on the GPU for the per-token case, but wanted to be on the CPU for the per-tensor case to avoid a synchronization.
• Having some tensors on the GPU and other tensors on the CPU breaks torch.compile
• Calling torch.item at all breaks CUDA graphs.

In order to fix this, I've further modified the CUTLASS epilogue code that loads and broadcasts the scales to always take a float*, and added a bool to indicate if the load should be broadcasting a vector or a scalar.

---

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!

[bnellnm]

LGTM

[pcmoritz]

The change looks good to me overall.

Before we merge this, I have an ask: Can you remove the spurious formatting changes of broadcast_load_epilogue_c2x.hpp in this PR and also for broadcast_load_epilogue_c3x.hpp compared to upstream? Keeping these files close to upstream will be crucial to make the code readable and incorporate any future changes from upstream. In particular, we should not be running clang format on these files and exclude them from clang-format.

[tlrmchlsmth]

@pcmoritz You are right -- I'll revert the formatting changes

[tlrmchlsmth]

Going to redo those files from upstream one more time since I'm seeing a bit of drift -- we should be rigorous about keeping track of what version of CUTLASS they come from, so I'll grab them from 3.5

[tlrmchlsmth]

I think this should be good to go now

[pcmoritz]

Nice, thanks for the modifications :)