Skip to content

bugfix: update xllm_ops submodule for dynamic mega chunk gdn heads.#1902

Merged
yingxudeng merged 2 commits into
xLLM-AI:mainfrom
Sinle4Cat:bugfix/update-xllm-ops-dynamic-gdn-main
Jul 10, 2026
Merged

bugfix: update xllm_ops submodule for dynamic mega chunk gdn heads.#1902
yingxudeng merged 2 commits into
xLLM-AI:mainfrom
Sinle4Cat:bugfix/update-xllm-ops-dynamic-gdn-main

Conversation

@Sinle4Cat

Copy link
Copy Markdown
Contributor

Description

Update the third_party/xllm_ops submodule to the latest xLLM-AI/xllm_ops main commit.

  • Included fix: dynamic mega_chunk_gdn synchronization update from xllm_ops.

This PR only changes the xllm_ops submodule gitlink. No xLLM source code is modified.

Related Issues

N/A

Change Type

  • Bug fix
  • New feature
  • Performance improvement
  • Refactor
  • Documentation
  • Test
  • Build or CI

Pull Request Checklist

Thank you for contributing to xLLM. Before requesting review, please make sure the following items are complete.

PR Title and Commit Messages

  • The PR title and each commit message follow the xLLM commit format: <type>: <subject>.

Allowed types: feat, bugfix, docs, test, refactor, chore, style, revert, perf, model, build, release.
The subject should use clear English, start with a verb, include at least 4 words, and end with ..

Pre-commit Checks

  • I have installed pre-commit by running pip install pre-commit or an equivalent command.
  • I have installed the hooks with pre-commit install.
  • I have run pre-commit run --all-files and fixed any reported issues.

If you are unsure how to set up pre-commit, see the pre-commit documentation.

Self Review

  • I have self-reviewed the code according to .agents/skills/code-review/references/custom-code-style.md, especially code written or assisted by AI.
  • I have rebased this PR onto the latest main branch.

Build and Test Coverage

  • Tests have been added or updated as needed.
  • CUDA: python setup.py build test has passed on a CUDA machine.
  • NPU: python setup.py build test has passed on an NPU machine.
  • MLU: python setup.py build test has passed on an MLU machine.

Reviewer Notes

  • This PR is a submodule pointer update only.
  • The target xllm_ops commit is the current xLLM-AI/xllm_ops main head: 7345ef6027167a6efd739370bee0a66a4c31ad6d.
  • xllm_ops verification on npu:
    • mega_chunk_gdn build passed.
    • test_mega_chunk_gdn.py: 28 passed.
    • Full xllm_ops custom ops build passed.
    • Python test files passed when run in isolated pytest processes; one single-process full pytest run had a transient order-
  • Qwen3.6-35B TP 1,2,4,8 benchmark passed on continuous NPU devices
  • Qwen3.6-35B no-MTP/MTP ceval dataset accuracy 91.68.
  • multimodal qwen3-5-27b TP 4:tested in textvqa accuracy gen 83.23

@yingxudeng yingxudeng left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

yingxudeng
yingxudeng previously approved these changes Jul 8, 2026
@Sinle4Cat Sinle4Cat dismissed stale reviews from weizhehuang0827 and yingxudeng via c0503d0 July 10, 2026 03:36
@Sinle4Cat Sinle4Cat force-pushed the bugfix/update-xllm-ops-dynamic-gdn-main branch from c0503d0 to ba2a48c Compare July 10, 2026 03:41
- xllm_ops -> 9e436ee (dynamic mega chunk gdn heads, laser_attention, LayerNorm op, CANN 9.0)
- xllm_atb_layers -> 67e56f2 (configurable scaleValue for aclnnXAttention, w4a8 quant for dsa)
@Sinle4Cat Sinle4Cat force-pushed the bugfix/update-xllm-ops-dynamic-gdn-main branch from ba2a48c to d3b9482 Compare July 10, 2026 03:46
@guojinrong-nn guojinrong-nn self-requested a review July 10, 2026 12:55
@yingxudeng yingxudeng merged commit 40b5ff5 into xLLM-AI:main Jul 10, 2026
10 of 16 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants