Skip to content

Enable DGX Spark (aarch64 + CUDA 13.0) support#2

Open
alexhegit wants to merge 1 commit into
NVlabs:mainfrom
alexhegit:spark
Open

Enable DGX Spark (aarch64 + CUDA 13.0) support#2
alexhegit wants to merge 1 commit into
NVlabs:mainfrom
alexhegit:spark

Conversation

@alexhegit

Copy link
Copy Markdown
  • Relax torch upper cap (<2.7 → no cap): CUDA 13.0 on aarch64 needs torch>=2.12
  • Relax torchvision upper cap (<0.22 → no cap): matching version needed
  • Make scene-synthesizer[recommend] x86_64-only: usd-core has no aarch64 wheel
  • Add AGENTS.md with platform notes table covering ptv3vanilla, pointnet2_ops, scene-synthesizer, spconv, and end2end compatibility on aarch64
  • Document ROS Jazzy pytest workaround

Verified: 30 CPU tests + 2 GPU inference tests pass on NVIDIA GB10.

- Relax torch upper cap (<2.7 → no cap): CUDA 13.0 on aarch64 needs torch>=2.12
- Relax torchvision upper cap (<0.22 → no cap): matching version needed
- Make scene-synthesizer[recommend] x86_64-only: usd-core has no aarch64 wheel
- Add AGENTS.md with platform notes table covering ptv3vanilla, pointnet2_ops,
  scene-synthesizer, spconv, and end2end compatibility on aarch64
- Document ROS Jazzy pytest workaround

Verified: 30 CPU tests + 2 GPU inference tests pass on NVIDIA GB10.
Signed-off-by: alexhegit <heye_dev@163.com>
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.

1 participant