Skip to content

Xpu/clean pr2 upstream sync - #3

Open
kahlun wants to merge 22 commits into
mainfrom
xpu/clean-pr2-upstream-sync
Open

Xpu/clean pr2 upstream sync#3
kahlun wants to merge 22 commits into
mainfrom
xpu/clean-pr2-upstream-sync

Conversation

@kahlun

@kahlun kahlun commented Jul 8, 2026

Copy link
Copy Markdown
Owner

What does this PR do?

Add concise overview of what this PR aims to achieve or accomplish. Reference related GitHub issues and PRs that help with the review.

Checklist Before Starting

  • Search for similar PRs. Paste at least one query link here: ...
  • Format the PR title as [{modules}] {type}: {description} (This will be checked by the CI)
    • {modules} include fsdp, megatron, veomni, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data, cfg, reward, fully_async, one_step_off
    • If this PR involves multiple modules, separate them with , like [megatron, fsdp, doc]
    • {type} is in feat, fix, refactor, chore, test
    • If this PR breaks any API (CLI arguments, config, function signature, etc.), add [BREAKING] to the beginning of the title.
    • Example: [BREAKING][fsdp, megatron] feat: dynamic batching

Test

For changes that can not be tested by CI (e.g., algorithm implementation, new model support), validate by experiment(s) and show results like training curve plots, evaluation results, etc.

API and Usage Example

Demonstrate how the API changes if any, and provide usage example(s) if possible.

# Add code snippet or script demonstrating how to use this

Design & Code Changes

Demonstrate the high-level design if this PR is complex, and list the specific changes.

Checklist Before Submitting

Important

Please check all the following items before requesting a review, otherwise the reviewer might deprioritize this PR for review.

kahlun and others added 5 commits July 7, 2026 19:23
Add comprehensive Intel GPU (Arc/Arc Pro B-series) support for veRL:

- Docker image with Intel compute-runtime 26.09, IGC 2.30.1, oneCCL 2021.15
- FSDP + vLLM rollout support for GRPO, PPO, and SFT algorithms
- Runtime patches for vLLM XPU platform and Intel GPU device detection
- Intel GPU-specific test scripts and environment configuration
- Documentation for Docker build and getting started guide

Validated on 2× Intel Arc Pro B60 (Battlemage):
- GRPO: 41.6 tok/s (2-GPU), 172 tok/s (1-GPU)
- PPO: 27.6 tok/s (2-GPU GAE with critic)
- SFT: Multi-GPU FSDP training with checkpointing

Co-authored-by: Claude <claude@anthropic.com>
Signed-off-by: Kah Lun Teoh <kah.lun.teoh@intel.com>
@kahlun kahlun closed this Jul 8, 2026
@kahlun kahlun reopened this Jul 8, 2026
kahlun added 17 commits July 7, 2026 21:06
…ne environment setup, and remove deprecated oneCCL workarounds
…ated environment scripts, and streamline memory reporting
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.

1 participant