[Cherry-pick] Port MoE multi-stream fix to v0.11.0-dev

[whx-sjtu]

This PR moves the communication operation of shared experts out of extra stream because I found that this might cause rtMemcpy related errors when running shared experts multistream with aclgraph.

Furthermore, I utilize a global variable as extra stream object to avoid allocating streams for each layer in full-graph mode.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

• A PR should do only one thing, smaller PRs enable faster reviews.
• Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
• Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

[gemini-code-assist]

Code Review

This pull request refactors the multi-stream handling for Mixture-of-Experts (MoE) layers, specifically for shared experts. The changes involve moving communication operations for shared experts out of the extra computation stream to prevent potential errors with aclgraph. Additionally, it introduces a global stream for shared expert calculations to avoid re-creating streams for each layer, which is a good optimization for graph mode. The implementation looks correct and aligns with the stated goals. I have one piece of feedback regarding a misleading comment in a new utility function, which should be corrected for clarity.