[main] Use AddRmsNormQuant ops in the custom model to optimize Qwen3's performance

[rjg-lyh]

What this PR does / why we need it?

Optimizes the performance of the Qwen3 quantization model by registering a custom model and adding the AddRmsNormQuant operation. Subsequent PRs will focus on performance optimizations based on this custom model.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

CI passed with existing test.

• vLLM version: v0.9.2
• vLLM main: vllm-project/vllm@8d0a01a

[codecov]

Codecov Report

Attention: Patch coverage is 49.38272% with 41 lines in your changes missing coverage. Please review.

Project coverage is 60.11%. Comparing base (bf25498) to head (bcbc024).
Report is 35 commits behind head on main.

Files with missing lines	Patch %	Lines
vllm_ascend/models/qwen3.py	48.43%	33 Missing ⚠️
vllm_ascend/ops/layernorm.py	27.27%	8 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1806      +/-   ##
==========================================
+ Coverage   54.93%   60.11%   +5.18%     
==========================================
  Files          80       74       -6     
  Lines        9712     8081    -1631     
==========================================
- Hits         5335     4858     -477     
+ Misses       4377     3223    -1154     

Flag	Coverage Δ
unittests	60.11% <49.38%> (+5.18%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[wangxiyuan]

do not rewrite the model arch, if the change is only AddRMSNormW8A8Quant.

From 0.9.2, vllm support custom ops override, we can register our ops when setup vllm-ascend, take #1647 for example.

[wangxiyuan]

As we talked offline, in the future, we should contribute the arch change to vLLM instead of maintain it inner vLLM Ascend. It's should be the common abiliity