[zxd1997066]

Overview

The purpose of this PR is to:

• Some quantization cases have OOM issue on B50 https://buildkite.com/vllm/intel-ci/builds/16237/list?jid=01a19ad8-298d-47b3-974d-419fa72eedd5&tab=output, increase the memory request to run them on B70
• Deselect one test case for the timeout issue in v1/shutdown https://buildkite.com/vllm/intel-ci/builds/16237/list?jid=01a19b25-b61e-4219-a746-3e554b7a5e3b&tab=output, An internal issue has been filed to track this.

Claims

Validation

label	agent	runtime in CI	estimated frequency (post merge only)
(BMG) Quantization	1xB70	11 min	~5.9 times/day

This doesn't put too much extra pressure on B70 resources.

Details

---

Pull Request Checklist

• I used vLLM's /pr-checklist skill. (Mandatory for agents, optional for humans).

• AI assistance was used during the creation of this PR.

• Design Fit: Minimizes impact on core components, reuses existing functionality, and justifies added complexity.

• Testing and Validation: Validates the change and ensures any added tests are meaningful and reliable, with CI coverage or documented CI resource constraints and validation performed outside CI.

• Code Quality and Style: Keeps code and comments clear and concise, and updates relevant documentation and examples.

• Pull Request Contents: Includes a brief summary and relevant links, supports claims with evidence, explains root causes and implementation trade-offs, and follows the contributing guide.

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[khluu]

Kevin's review agent (Claude Code) here.

The memory bump in quantization_intel.yaml looks good. One question before I stamp the CI changes: this PR also adds a deselect to distributed_intel.yaml:

--deselect="tests/v1/shutdown/test_forward_error.py::test_llm_model_error[hmellor/tiny-random-LlamaForCausalLM-2-True]"

That turns off the TP=2 case of the engine-dies-on-model-error test on XPU, but the title and description only mention the Quantization memory change. Could you explain why it's being skipped (failure link or tracking issue), add that to the description, or move it to its own PR? Happy to approve once that's clarified.

[zxd1997066]

Kevin's review agent (Claude Code) here.

The memory bump in quantization_intel.yaml looks good. One question before I stamp the CI changes: this PR also adds a deselect to distributed_intel.yaml:

--deselect="tests/v1/shutdown/test_forward_error.py::test_llm_model_error[hmellor/tiny-random-LlamaForCausalLM-2-True]"

That turns off the TP=2 case of the engine-dies-on-model-error test on XPU, but the title and description only mention the Quantization memory change. Could you explain why it's being skipped (failure link or tracking issue), add that to the description, or move it to its own PR? Happy to approve once that's clarified.

Thanks for your review. I have updated the title and description.

[zxd1997066]

/ci run

[github-actions]

✅ Triggered Buildkite CI #99435 for commit bcf19f66a3f5.