Search before asking
Question
Hello,
I have a Jetson Thor running R38 (Ubuntu 24.04 host) and want to run Autoware for system-level profiling (end-to-end latency, ROS 2 throughput, etc.).
These are the problems I am facing:
Autoware targets Ubuntu 22.04 + ROS 2 Humble
L4T containers for R38 are based on Ubuntu 24.04, which is needed for GPU access on Thor
Generic Ubuntu 22.04 containers won’t have a matching CUDA/TensorRT user-space for the R38 driver stack, so CUDA doesn't work
The questions I have are:
Is there a ROS 2 Humble container available or planned for the R38 stack?
Has anyone successfully run Autoware on Jetson Thor?
What would be a recommended approach for this use case, and given system setup?
Additional
No response
Search before asking
Question
Hello,
I have a Jetson Thor running R38 (Ubuntu 24.04 host) and want to run Autoware for system-level profiling (end-to-end latency, ROS 2 throughput, etc.).
These are the problems I am facing:
Autoware targets Ubuntu 22.04 + ROS 2 Humble
L4T containers for R38 are based on Ubuntu 24.04, which is needed for GPU access on Thor
Generic Ubuntu 22.04 containers won’t have a matching CUDA/TensorRT user-space for the R38 driver stack, so CUDA doesn't work
The questions I have are:
Is there a ROS 2 Humble container available or planned for the R38 stack?
Has anyone successfully run Autoware on Jetson Thor?
What would be a recommended approach for this use case, and given system setup?
Additional
No response