Support Matrix#

This matrix describes the supported AutoData v0.1.0 stack. Upstream projects may support additional platforms, but configurations outside this matrix are not validated by AutoData. Use the recursively pinned submodules from the AutoData repository instead of independently selecting Isaac Lab or Isaac Lab-Arena revisions.

AutoData platform and resource support#

Area

Supported or required configuration

Host operating system

Linux x86_64; Ubuntu 22.04 or 24.04

GPU

NVIDIA RTX GPU with RT cores and at least 16 GB VRAM. See Isaac Sim 6.0.1 requirements.

NVIDIA driver

Linux 595.58.03. See Isaac Sim 6.0.1 requirements.

CPU and RAM

Minimum: 4 CPU cores and 32 GB RAM. Recommended: 8 or more cores and 64 GB RAM. See Isaac Sim 6.0.1 requirements.

Disk

Minimum: 100 GB SSD. Recommended: 500 GB SSD

Docker installation

Docker Engine with the NVIDIA Container Toolkit; recommended AutoData installation

Conda installation

Optional Linux installation using conda, uv, and Python 3.12

Isaac Sim

6.0.1 container image; 6.0.1.0 Python package

Isaac Lab

3.0.0 at commit ffff603eafc6b74264a5261cc0183d6a65390d78

Isaac Lab-Arena

0.2.0 at commit 8b82dca224f2b5af08f339f987613c59ce9cdbaa

Python and PyTorch

Python 3.12 and PyTorch 2.10.0 with CUDA 12.8

Git and Git LFS

Git with recursive submodule support and Git LFS

cuRobo and SkillGen

Optional; cuRobo commit ebb71702f3f70e767f40fd8e050674af0288abe8 and CUDA Toolkit 12.8