Add context parallel config options

[adibarra]

Summary

Adds first-class single-node master-config options for vLLM context parallelism:

• dcp-size
• pcp-size

Both default to 1, preserving existing behavior.

These values now flow from master YAML through matrix generation and workflows as DCP_SIZE / PCP_SIZE, removing the need for concurrency-dependent Bash logic such as enabling DCP only when CONC >= 16.

The DSv4 B200/B300 agentic vLLM recipes translate them to:

• --decode-context-parallel-size
• --prefill-context-parallel-size

DCP reuses TP GPUs. PCP uses TP × PCP_SIZE GPUs, so allocation and per-GPU throughput accounting are updated accordingly. Result identities also include DCP/PCP to avoid collisions or incorrect artifact reuse.

Multinode/disaggregated worker configuration and production tuning rows are unchanged.

Verification

• 293 focused tests passed
• Modified workflows parsed successfully
• Modified shell scripts passed bash -n

[github-actions]

Thanks for the contribution! Please reach out to respective companies' CODEOWNER to fill in the latest PR_REVIEW_CHECKLIST.md before pinging core maintainer on Slack for review. In order for the signoff PR check bot to trigger, you must follow the PR_REVIEW_CHECKLIST.md template correctly, including the phrase As a PR reviewer and CODEOWNER, I have reviewed this and have.

For PR verification, add the full-sweep-fail-fast label (strongly recommended) to this PR — the benchmark sweep only runs on labeled PRs. Use full-sweep-enabled only if you need matrix jobs to keep running past a failure.

PR authors are responsible for ensuring that after merging, all GitHub Action jobs fully pass. A lot of the time, failures are just flakes and simply re-running the failed jobs will fix it. See GitHub's docs on re-running failed jobs

---

感谢你的贡献！请联系相应公司的 CODEOWNER 填写最新的 PR_REVIEW_CHECKLIST.md，然后再在 Slack 上联系核心维护者进行审阅。为了触发 signoff PR 检查机器人，你必须正确遵循 PR_REVIEW_CHECKLIST.md 模板，包括保留英文语句 As a PR reviewer and CODEOWNER, I have reviewed this and have。

如需进行 PR 验证，请为此 PR 添加 full-sweep-fail-fast 标签（强烈推荐）— 基准测试 sweep 仅在带有标签的 PR 上运行。仅当需要矩阵任务在失败后继续运行时才使用 full-sweep-enabled。

PR 作者有责任确保合并后所有 GitHub Action 任务完全通过。 很多时候失败只是偶发抖动（flake），重新运行失败的任务即可解决。参见 GitHub 关于重新运行失败任务的文档