[Perf][Qwen3-TTS] Vectorize _LocalPredictorKVCache to eliminate per-token loops

[gcanlin]

PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED.

Purpose

Related to #938.

• Vectorize block table initialization: replace nested Python for i/for j loop with a broadcasted outer product of two torch.arange tensors
• Vectorize positions computation in build_attn_metadata: replace per-request torch.arange + list append + torch.cat loop with a 2D grid + boolean mask approach
• Vectorize slot_mapping computation: replace doubly-nested per-token Python loop (with .item() calls on every token) with bulk torch.repeat_interleave, integer division, and advanced indexing

Test Plan

To be done.

Test Result

---

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. Please provide the test scripts & test commands. Please state the reasons if your codes don't require additional test scripts. For test file guidelines, please check the test style doc
• The test results. Please paste the results comparison before and after, or the e2e results.
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model. Please run mkdocs serve to sync the documentation editions to ./docs.
• (Optional) Release notes update. If your change is user-facing, please update the release notes draft.

BEFORE SUBMITTING, PLEASE READ https://github.com/vllm-project/vllm-omni/blob/main/CONTRIBUTING.md (anything written below this line will be removed by GitHub Actions)