DSR1-FP4 MI355x SGLang: Add EP configurations

[ppalanga]

This PR does two things for DSR1-FP4 SGLang for MI355x:

1. Adding EP configurations
2. Update the sglang image to 0.5.13

---

Note

Low Risk
Benchmark and CI config only; no production serving or security paths. Main operational impact is longer/more expensive GPU sweeps from higher concurrency and new EP matrix points.

Overview
Extends dsr1-fp4-mi355x-sglang in amd-master.yaml for DeepSeek-R1 FP4 on MI355X: SGLang image v0.5.12 → v0.5.13, fixed-seq-len concurrency sweeps conc-end 64 → 256 for TP-only points, and new expert-parallel search-space rows (TP4/EP4 and TP8/EP8) for both 1k/1k and 8k/1k scenarios.

dsr1_fp4_mi355x.sh now requires EP_SIZE, passes --ep-size when EP > 1, and raises --cuda-graph-max-bs from 128 → 512 to match the wider concurrency range.

Documents the change under dsr1-fp4-mi355x-sglang in perf-changelog.yaml.

^{Reviewed by Cursor Bugbot for commit 008eeaa. Bugbot is set up for automated code reviews on this repo. Configure here.}

[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://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node 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/sglang/tree/main/docs_new

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.

[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://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node 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/sglang/tree/main/docs_new

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.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 016686f. Configure here.}