port MI355X DeepSeek V4 ATOM disagg to srt-slurm

[cquil11]

Summary

Ports the active dsv4-fp4-mi355x-atom-disagg 8k/1k submission from the configuration-specific AMD launcher to the ATOM/Infera implementation in SemiAnalysisAI/srt-slurm#3.

Scope

• preserve the original 1P1D TP8 points: c4, c8, c16, c32, c64, c128
• preserve the original 2P1D DPA+TP8 points: c256, c512, c768, c1024, c2048
• run the unchanged InferenceX utils/bench_serving/benchmark_serving.py through the custom benchmark contract
• use Infera discovery and KV-aware routing with ATOM KV events
• use the hardware-proven Mooncake TCP path initially; RDMA validation follows with the updated ATOM/Mooncake runtime
• remove the exact legacy wrapper and obsolete deprecated 1k/1k entry that depended on it

Dependency structure

• base InferenceX ATOM/Infera validation: feat: validate ATOM and Infera on MI300X and MI355X #2558
• srt-slurm ATOM/Infera implementation: feat: add ROCm ATOM backend and Infera routing srt-slurm#3
• Infera KV-event decoding fix: fix(router): decode ATOM KV events with map schema AMD-AGI/Infera#114
• ATOM Mooncake TCP fix: fix(mooncake): keep TCP transport off RDMA devices ROCm/ATOM#1866

Local validation

• both recipes pass srtctl dry-run
• generated InferenceX matrix contains exactly two topology jobs
• master config validation passes
• 171 InferenceX matrix/runner/gating tests pass
• 13 targeted srt-slurm AMD/ATOM/Infera tests pass
• YAML parsing, shell syntax, and git diff --check pass

Hardware validation

WIP. Exact workflow links and Slurm/artifact evidence will be added here as the full sweep runs.

[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

[cquil11]

Live 2P/1D DPA validation update:\n\n- Slurm 39443 completed all 64 model shards and CUDA-graph capture on both prefills and the decode.\n- Startup then failed deterministically in every logical worker because each ATOM data-parallel EngineCore tried to bind the same KV-event ZMQ endpoint. One DP rank bound successfully; the others raised Address already in use from Infera BlockManager hook initialization, so the worker could not register.\n- This is an Infera multi-DP KV-event publisher bug, not Slurm, MI355X hardware, model loading, Mooncake, or the srt-slurm topology.\n- The upstream fix is open as AMD-AGI/Infera#121. It gives every DP rank a local PUB ingress and relays all rank streams through the one endpoint advertised for the logical worker, retaining complete KV-aware cache visibility.\n- Focused engine/router validation is green: 50 tests passed, including two simultaneous EngineCore publishers through one advertised endpoint; Ruff, pre-commit, and diff checks passed.\n- The current exact-head workflow was not cancelled. Its 1P/1D arm is still running naturally as Slurm 39448. I will pin the reviewed fix and launch a fresh exact-head validation only after this attempt closes.