Adding new sharding and performance args for vLLM

[NicoGrande]

Description

This PR adds new flags to enable attn_dp in the vLLM mesh, as well as flags to tune the performance of vLLM generation. More information on these flags and their effects can be found here:
https://docs.vllm.ai/en/stable/configuration/optimization/#performance-tuning-with-chunked-prefill

Tests

Tested e2e RL workflow with new flags and updated Tunix code.

Checklist

Before submitting this PR, please make sure (put X in square brackets):

• I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have run end-to-end tests tests and provided workload links above if applicable.
• I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[richjames0]

LGTM