Skip to content

[AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260926 - #3364

Closed
yichiche wants to merge 3 commits into
mainfrom
amd/qwen35-agentx-image-20260922
Closed

[AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260926#3364
yichiche wants to merge 3 commits into
mainfrom
amd/qwen35-agentx-image-20260922

Conversation

@yichiche

@yichiche yichiche commented Sep 22, 2026 •
edited
Loading

Copy link
Copy Markdown
Collaborator

Summary

  • Bump qwen3.5-fp4-mi355x-sglang-agentic-mtp from lmsysorg/sglang-rocm:v0.5.19-rocm720-mi35x-20260915 to lmsysorg/sglang-rocm:v0.5.20-rocm720-mi35x-20260926 (Docker Hub tag HTTP 200, digest sha256:f71b2a13a23132b9efdd39df58d13f310a59c2d101ec72271457295ed48016a9). The 20260925 nightly was never published.
  • Keep TP4 GPU-resident conc [1, 4, 8, 12, 16]. Move TP2 conc 16 onto HiCache and drop TP2 GPU-resident conc 20. TP2 resident is [1, 4, 8, 12]; TP2 HiCache is [16, 20, 24, 28, 32, 36, 40].
  • Pin the HiCache host pool with hicache-size: 253 (overrides hicache-ratio) on every TP2 HiCache point. Other recipe flags are unchanged.

Details

AgentX on main now runs through native srt-slurm, and the legacy qwen3.5_fp4_mi355x_sglang_mtp.sh script was removed. This PR therefore changes:

  • configs/amd-master.yaml: the image and the TP2 conc grids for the AgentX arm.
  • benchmarks/single_node/srt-slurm-recipes/qwen3.5/sglang/mi355x-fp4-mtp/agentic.yaml: model.container set to the same image; override_tp2_c16 and override_tp2_c20 (resident) removed; new override_tp2_c16_hicache added; hicache-size: 253 added to every HiCache override.
  • perf-changelog.yaml: one appended entry.

Each of the 16 matrix points resolves to exactly one recipe override under validate_recipe (gpus, CONC, KV_OFFLOADING and image all match), and no override is left unused. Fixed-seq MI355X SGLang arms are a separate PR.

AI model disclosure

Cursor Grok 4.6 prepared the image bump, HiCache size pin, changelog, and this PR text.

@yichiche
yichiche requested a review from a team September 22, 2026 13:52
@github-actions

github-actions Bot commented Sep 22, 2026

Copy link
Copy Markdown
Contributor

Thanks for the contribution!

  • Review: If this PR changes files owned by someone other than a repository admin or @SemiAnalysisAI/core, ask one eligible CODEOWNER to complete the latest PR_REVIEW_CHECKLIST.md before contacting a core maintainer on Slack. Follow the template exactly, including As a PR reviewer and CODEOWNER, I have reviewed this and have, so sign-off verification triggers.
  • PR verification: Sweeps only run on labeled PRs. Add full-sweep-fail-fast (strongly recommended); use full-sweep-enabled only when matrix jobs should continue after a failure.
  • After merging: PR authors must ensure all GitHub Actions jobs pass. Transient failures often pass on rerun; see how to rerun failed jobs.
中文

感谢你的贡献！

  • **审阅：**如果 PR 修改的文件归属于仓库管理员及 @SemiAnalysisAI/core 之外的 CODEOWNER，请先联系一位有资格的 CODEOWNER 填写最新的 PR_REVIEW_CHECKLIST.md，再通过 Slack 联系核心维护者。必须严格遵循模板，并保留 As a PR reviewer and CODEOWNER, I have reviewed this and have，才能触发签核验证。
  • **PR 验证：**扫描仅在带有标签的 PR 上运行。强烈建议添加 full-sweep-fail-fast；仅当需要矩阵任务在失败后继续运行时才使用 full-sweep-enabled。
  • **合并后：**PR 作者必须确保所有 GitHub Actions 任务通过。临时性失败通常可以通过重新运行恢复；参见重新运行失败任务的说明。

@yichiche
yichiche force-pushed the amd/qwen35-agentx-image-20260922 branch from 6e4e237 to 79cd4b8 Compare September 22, 2026 13:52
@yichiche yichiche added agentx AgentX benchmarks, recipes, and infrastructure and removed agentx AgentX benchmarks, recipes, and infrastructure labels Sep 22, 2026
@github-actions

This comment was marked as outdated.

Sign in to view
claude[bot]

This comment was marked as outdated.

Sign in to view

@yichiche yichiche changed the title [AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260922 [AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260924 Sep 24, 2026
@yichiche
Merge remote-tracking branch 'origin/main' into amd/qwen35-agentx-ima…
7ba1a8d 
…ge-20260922

# Conflicts:
#	perf-changelog.yaml
@yichiche yichiche added full-sweep-enabled agentx AgentX benchmarks, recipes, and infrastructure and removed full-sweep-fail-fast agentx AgentX benchmarks, recipes, and infrastructure full-sweep-enabled labels Sep 24, 2026
@functionstackx

functionstackx commented Sep 26, 2026

Copy link
Copy Markdown
Collaborator

InferenceX has switched away from unmaintainable bash scripts to YAML files that don't repeat the same stuff over and over again. Please merge the latest main into this PR: we have migrated single-node AgentX onto native srt-slurm (#3428), so AgentX configs are now declarative YAML recipes, not per-config 1000+ line bash slop scripts. Please also delete the old benchmarks/single_node/** scripts (see this recipe for the new format).

@functionstackx functionstackx mentioned this pull request Sep 26, 2026
Open
1 task
functionstackx added a commit that referenced this pull request Sep 26, 2026
@functionstackx @claude
Port #3364 TP2 conc grid and hicache-size 253 onto the srt-slurm recipe
afc4a0b 
TP2 GPU-resident KV now covers [1, 4, 8, 12] and TP2 HiCache
[16, 20, 24, 28, 32, 36, 40]; every HiCache variant pins
hicache-size 253. Drops the now-unused tp2_c16/tp2_c20 resident
variants and adds tp2_c16_hicache so each matrix point resolves to
exactly one recipe variant.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@functionstackx

functionstackx commented Sep 26, 2026

Copy link
Copy Markdown
Collaborator

Closing in favor of #3455, which ports this PR's changes onto the srt-slurm recipe that #3428 switched qwen3.5-fp4-mi355x-sglang-agentic-mtp to: the v0.5.20-rocm720-mi35x-20260924 image, TP2 resident conc [1, 4, 8, 12], TP2 HiCache conc [16, 20, 24, 28, 32, 36, 40], and hicache-size: 253 on every HiCache variant. The legacy benchmarks/single_node/agentic/qwen3.5_fp4_mi355x_sglang_mtp.sh script this PR edits is no longer used by that config. Thanks for the tuning!

@yichiche yichiche changed the title [AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260924 [AMD][Qwen3.5] Bump MI355X AgentX image to v0.5.20-rocm720-mi35x-20260926 Sep 26, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@claude claude[bot] claude[bot] left review comments
@billishyahao billishyahao Awaiting requested review from billishyahao billishyahao is a code owner
@chunfangamd chunfangamd Awaiting requested review from chunfangamd chunfangamd is a code owner
@seungrokj seungrokj Awaiting requested review from seungrokj seungrokj is a code owner
@yctseng0211 yctseng0211 Awaiting requested review from yctseng0211 yctseng0211 is a code owner
@1am9trash 1am9trash Awaiting requested review from 1am9trash 1am9trash is a code owner

Assignees

No one assigned

Labels

full-sweep-enabled

Projects

InferenceX Board
Status: Done

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@yichiche @functionstackx @ChangLiu0709