Passing GPU of jetson xavier nx to container

Autonomous Machines Jetson & Embedded Systems Jetson Xavier NX
1

how do i pass the GPU my containers? every documented example i see refers to discrete gpu where nvidia-smi works.

i have a vanila install from sdk manager.

3

Hi @joel.anthony1, you just have to use --runtime nvidia when you start the container with docker run / ect. On Jetson there is no nvidia-smi, regardless of container or if GPU is enabled in container.

When you say “vanilla” install from SDK Manager, that’s assuming it installed the NVIDIA Container Runtime for you. Check that it’s listed under docker info:

sudo docker info | grep nvidia
 Runtimes: io.containerd.runc.v2 io.containerd.runtime.v1.linux nvidia runc
5

Runtimes: io.containerd.runc.v2 io.containerd.runtime.v1.linux nvidia runc
Default Runtime: nvidia

6

–runtime Nvidia should expose the underlying hardware to my container?

7

Yes, although it should use l4t-base or l4t-jetpack, ect as base container - you can find the instructions for running a container here (l4t-base, for example)

9

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.