meta-pytorch / meta-pytorch/torchcodec

Request for ROCM backend support

Open
#444 7 comments 10 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement module: rocm
Dominant language
Python
Stars
1.2k
Forks
125
Avg merge
22h 47m
Merged PRs (30d)
54

Description

The feature

Are there any plans on supporting ROCm backends? The current implementation relies on CUDA libraries libnpp and libnvrtc which are, to the extent of my knowledge, unavailable for direct use on ROCm devices. Are there any plans or timeline for an extension to support ROCm based GPU's?

Motivation, pitch

ROCm is getting more and more attention as the software continues to incrementally get better, and any user on the ROCm stack could get the same benefits to that of CUDA users if the codec ends up supporting ROCm.

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

The issue names no files, tests, or entry points; start by locating the CUDA-dependent libnpp and libnvrtc integration. Define the supported ROCm backend scope and validate it with backend tests, since the issue currently asks for plans rather than a specific change.

Written by the indexing model from the issue text.

Assessment

Domain
backend
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.