Skip to content

Add nvidia.com-cuda devcontainer feature - #87

Merged
sebst merged 3 commits into
mainfrom
copilot/add-nvidia-com-cuda-feature
Mar 16, 2026
Merged

sebst merged 3 commits into
mainfrom
copilot/add-nvidia-com-cuda-feature

Conversation

Copilot AI commented Mar 16, 2026 •

Copy link
Copy Markdown
Contributor

Adds a new community devcontainer feature that installs the NVIDIA CUDA Toolkit via NVIDIA's official apt repository.

Feature: nvidia.com-cuda

  • src/nvidia.com-cuda/devcontainer-feature.json — Feature manifest with a version option accepting "latest" or "X.Y" format (e.g. "12.6"); proposals cover CUDA 11.8–12.6
  • src/nvidia.com-cuda/install.sh — Detects distro (Ubuntu 20.04/22.04/24.04, Debian 11/12) and arch (x86_64, aarch64→sbsa), installs the NVIDIA CUDA apt keyring, then installs cuda-toolkit or cuda-toolkit-X-Y for a pinned version
  • test/nvidia.com-cuda/test.sh — Verifies nvcc --version succeeds post-install
  • .github/workflows/test.yaml — Adds nvidia.com-cuda to both test-autogenerated and test-scenarios matrix jobs

Usage

// devcontainer.json
{
  "features": {
    "ghcr.io/devcontainer-community/features/nvidia.com-cuda:1": {
      "version": "12.6"   // or "latest"
    }
  }
}

The keyring URL is parameterised via cuda_keyring_version so bumping the keyring package requires a one-line change.

Original prompt

This section details on the original issue you should resolve

<issue_title>Add feature nvidia.com-cuda</issue_title>
<issue_description></issue_description>

Comments on the Issue (you are @copilot in this section)


💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

Co-authored-by: sebst <592313+sebst@users.noreply.github.com>
Copilot AI changed the title [WIP] Add feature nvidia.com-cuda Add nvidia.com-cuda devcontainer feature Mar 16, 2026
Copilot AI requested a review from sebst March 16, 2026 15:04
@sebst

sebst commented Mar 16, 2026

Copy link
Copy Markdown
Member

@copilot Tests for nvidia.com-cuda are failing:

🔄 Testing 'nvcc installed'

bash: line 1: nvcc: command not found`

Co-authored-by: sebst <592313+sebst@users.noreply.github.com>

Copilot AI commented Mar 16, 2026

Copy link
Copy Markdown
Contributor Author

@copilot Tests for nvidia.com-cuda are failing:

...

Fixed in commit 5805b01. The CUDA toolkit installs nvcc to /usr/local/cuda/bin/ which isn't in the default PATH. Two changes in install.sh:

  1. Creates /etc/profile.d/cuda.sh prepending /usr/local/cuda/bin to PATH for interactive shells
  2. Symlinks /usr/local/bin/nvcc → /usr/local/cuda/bin/nvcc so nvcc is immediately available in non-interactive shells (including the test runner)

@sebst
sebst marked this pull request as ready for review March 16, 2026 15:44
@sebst
sebst merged commit a34e57f into main Mar 16, 2026
93 checks passed
@sebst
sebst deleted the copilot/add-nvidia-com-cuda-feature branch March 23, 2026 19:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add feature nvidia.com-cuda

2 participants