Fix 310P issues in main

[leo-pony]

What this PR does / why we need it?

Fix 310P issues in main.

Test on 310p NPU host, models as following:
Qwen/Qwen3-0.6B
Qwen/Qwen2.5-7B-Instruct
Qwen2.5-VL-3B-Instruct
Test cases are passed.

Know Issue:
For pangu-pro-moe-model model, has error:
ValueError: Unsupported soc_version: AscendSocVersion.UNDEFINED
Detail can see in PR:#2676

Does this PR introduce any user-facing change?

How was this patch tested?

• vLLM version: v0.11.0rc3
• vLLM main: vllm-project/vllm@c9461e0

[gemini-code-assist]

Code Review

This pull request introduces fixes for the 310P platform. The changes include correcting the model used in a vision-language test, adding a dedicated KV cache initialization path for 310P, and skipping ATB warmup on this platform.

My review identifies a couple of issues in the new _initialize_kv_cache_tensors_310p method. There's a critical bug where the KV cache tensors are created with a hardcoded torch.float16 dtype, which should be corrected to use the dtype from the kv_cache_spec. Additionally, there's a misleading error message when KV cache transfer is used, which could make debugging difficult. I've provided suggestions to fix both of these issues.

[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.

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.