Open xxRockOnxx opened 2 years ago
@xxRockOnxx could you uncomment the #debug =
lines in /etc/nvidia-container-runtime/config.toml
repeat one of the tests and attach the contents of /var/log/nvidia-container-toolkit.log
and (if it exists -- depending on the test) /var/log/nvidia-container-runtime.log
?
Here's the requested log:
For context:
Model: GTX 1660 Ti Driver version: 510.60.02 CUDA version: 11.6 nvidia-container-toolkit version: 1.9.0 nvidia-container-runtime version: 3.9.0
OS: Arch Kernel: 5.17.3-arch1-1
When running the following commands:
docker run --rm --gpus all nvidia/cuda:11.6.2-runtime-ubuntu20.04 nvidia-smi
docker run --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=all nvidia/cuda:11.6.2-runtime-ubuntu20.04 nvidia-smi -L
docker run --rm --gpus '"device=0"' --device /dev/nvidia-caps --device /dev/nvidia0 --device /dev/nvidiactl --device /dev/nvidia-modeset --device /dev/nvidia-uvm --device /dev/nvidia-uvm-tools nvidia/cuda:11.0.3-base-ubuntu20.04 nvidia-smi
gives no output. It's hard to figure what's happening because even an error isn't being given.
I tried the following:
systemd.unified_cgroup_hierarchy=false
no-cgroups = false
in/etc/nvidia-container-runtime/config.toml
user
in/etc/nvidia-container-runtime/config.toml
as mentioned in https://github.com/NVIDIA/nvidia-docker/issues/850/etc/docker/daemon.json
as mentioned in https://github.com/NVIDIA/nvidia-container-toolkit/issues/12Related issues:
Guide tried: https://forum.manjaro.org/t/howto-installing-docker-and-nvidia-runtime-my-experience-and-howto/97017
I'm ran out of things to try at this point.