NVIDIA / NVIDIA/cuEquivariance

Trying to run boltz2 on AMD GPUS

Open
#294 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
433
Forks
42
PR merge metrics
No merged PRs in 30d

Description

We are trying to run our boltz2 on AMD series accelerators. However, the project currently only provides pre‑built packages (cuequivariance-ops-cu12 and cuequivariance-ops-torch-cu12) for NVIDIA CUDA. Since our workflow relies on the source code of cuequivariance-ops (which is tightly coupled to CUDA‑specific kernels), we are unable to adapt it to a HIP/ROCm environment.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing the cuequivariance-ops source and the pre-built cuequivariance-ops-cu12 and cuequivariance-ops-torch-cu12 packages mentioned in the issue, focusing on their CUDA-specific kernels. Determine the scope needed for HIP/ROCm support and validate whether boltz2 can run on AMD accelerators without the CUDA-only packages.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.