support silu vectorization

[stickingjh]

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.

Support silu vectorization if d is multiple of vec size, later odd dim vectorization will be supported. The support is similar to flashinfer support. From local testing, perf is improved by about 60-70%.

(Optional) Documentation Update

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Code Review

This pull request introduces a vectorized CUDA kernel for SiLU activation to improve performance when the tensor dimension is a multiple of the vector size. The changes include the new kernel, a dispatch mechanism to select the appropriate kernel, and updated tests to cover both the vectorized and non-vectorized paths. My review identified two critical issues concerning memory alignment in the new CUDA kernel. Correcting these will ensure the stability and correctness of the vectorized implementation. Once these are addressed, this will be a solid performance improvement.

[LucasWilkinson]

why is the griddepcontrol needed? doesnt seem to overlapping anything