[WIP] Updated DSv4 vllm B300 MTP

[wzhao18]

No description provided.

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://cookbook.sglang.io/
• https://docs.vllm.ai/projects/recipes/en/latest/

If it is not, please create a PR first before we can merge your PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sgl-cookbook

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. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[claude]

Claude finished @wzhao18's task in 1m 28s —— View job

---

Review of PR #1271

• Gather context and read changed files
• Validate master config changes
• Validate perf-changelog.yaml
• Validate benchmark script

Summary

LGTM - no blocking issues found.

What this PR does: Upgrades the vLLM image from v0.20.0-cu130 to v0.20.1-cu130 for the B300 MTP config and adds new data-parallel attention + expert parallelism search space entries. The benchmark script is updated to conditionally set --moe-backend deep_gemm_mega_moe and MAX_NUM_BATCHED_TOKENS=2048 when DP_ATTENTION=true.

Checks passed:

• Image vllm/vllm-openai:v0.20.1-cu130 is publicly accessible on Docker Hub
• Model prefix dsv4 is valid
• perf-changelog.yaml entry is correctly appended at the end of the file
• Expert parallelism is correctly conditional on EP_SIZE (pre-existing)
• --dsv4 flag implies --use-chat-template (verified in benchmark_lib.sh)
• Server launch command is properly formatted with line continuations

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25297833177
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25297833177

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25300229552
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25300229552