MicroOS - Immich with CUDA/NVIDIA?

English Hardware
1

Hi,

I’m successfully running Immich on my MicroOS server using rootless Podman and quadlets, however, I’m going through the machine-learning guide/hardware acceleration at https://docs.immich.app/features/ml-hardware-acceleration/ and it mentions I need to have the NVIDIA Container Toolkit installed?

I noticed that this wasn’t recommended in the OpenSUSE NVIDIA Driver page, and also that is isn’t available in MicroOS anyway…

Is there any way I can get CUDA support in MicroOS, or am I missing something obvious?

Thanks for any help!

2

@Shortwave8115 So add the nvidia container repository? If using the container the only thing needed on the host is the nvidia kmp no other libraries or packages.

zypper ar https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo
zypper mr -gf nvidia-container-toolkit
zypper in -r nvidia-container-toolkit nvidia-container-toolkit

So prefix transactional-update run to the above, then either reboot or transactional-update apply

3

I see the docs mention not to use the toolkit, but I’m not sure why. https://en.opensuse.org/SDB:NVIDIA_drivers#CUDA

I see there’s other steps here for adding CUDA in MicroOS, I’ll take a look at these too.

4

Follow the upstream instructions… “On Linux (except for WSL2), you also need to have NVIDIA Container Toolkit installed.”

I use the container here, but on Leap 16.0 and for kubernetes.

It needs the container to run a docker container, you won’t get that by installing cuda on the host…