[Bugfix] make moe_align_block_size compliable with cuda graph

[jinzhen-lin]

The moe_align_block_size is used by many moe models (e.g. DeepSeek-V3) but it is not compliable with cuda graph now. This PR fix it.

Reference: sgl-project/sglang@77d1210 sgl-project/sglang@6e53051

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
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 do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[jinzhen-lin]

With cuda graph, the generation speed of DeepSeek-V3 (W4A16, 8*A100-80G, bs=1) increase from 5 tokens/s to 10 tokens/s.

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @jinzhen-lin.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[jinzhen-lin]

close this PR as I create a new PR with better moe_align_block_size, see #12222