[Bugfix] fix an illegal memory access was encountered of marlin kernel + act_order

[jinzhen-lin]

Fix #18147

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[mgoin]

Thanks for the fix, LGTM! Let's see if CI is green

[jinzhen-lin]

@mgoin The failed tests seems unrelated. Quantization test also failed on main branch. https://buildkite.com/vllm/ci/builds/20197#0196d740-afa9-4124-acdf-c971cdf483f2

[DarkLight1337]

This PR is breaking LoRA Tests 4. It wasn't failing on the previous PR on main

[jinzhen-lin]

This PR is breaking LoRA Tests 4. It wasn't failing on the previous PR on main

It seems that the LoRA tests are passed after retry https://buildkite.com/vllm/ci/builds/20254 . This PR does nothing to lora, and the previous failed model is not a quantized model, so it may not related to this PR.

[DarkLight1337]

I see, thanks for the info. I'll keep an eye out for further failres then.

[DarkLight1337]

On the other hand, a bunch of distributed tests are failing on this PR. They are still failing after retrying them.

[jinzhen-lin]

On the other hand, a bunch of distributed tests are failing on this PR. They are still failing after retrying them.

Very strange. All failed models are not quantized models. And those failed tests are passed again after two commits https://buildkite.com/vllm/ci/builds/20259 . No idea now.