Hello there,
I couldn’t find a space or place to ask about where to discuss the gfx1030 pr feat(rocm): enable gfx1030 (RDNA2) as a recognized RDNA arch by BlivionIaG · Pull Request #52391 · vllm-project/vllm · GitHub
ROCm has been supporting gfx1030 for a while and it works when manually building, I have some hip kernels for rdna2 in testing as well. It would be nice to have it as a target for rocm builds.
Regards,
BlivionIaG
The best place to discuss the gfx1030 PR (feat(rocm): enable gfx1030 (RDNA2) as a recognized RDNA arch by BlivionIaG · Pull Request #52391 · vllm-project/vllm · GitHub) is directly in the PR’s comment thread on GitHub. The vLLM team and contributors actively monitor PR discussions, and reviewers provide feedback there. You can also join the vLLM developer Slack at https://slack.vllm.ai for real-time discussion in channels like #pr-reviews or #feat-rocm, as mentioned in the PR comments and vLLM contribution guidelines.
If you have specific technical questions or want to coordinate on testing/merging, commenting on the PR itself is encouraged and will reach the relevant maintainers. For broader roadmap or feature discussions, the Slack workspace is recommended. According to the vLLM contribution docs, you may also email the team for expedited review if needed, but for technical PRs, GitHub and Slack are the primary venues for discussion and collaboration. Would you like more detail on the review or merge process for ROCm-related PRs?
Sources:
Hint: Mention @Herald in the post for followups.