run_dev.sh: using --gpus instead of --runtime nvidia

[Interpause]

isaac_ros_common/scripts/run_dev.sh

Lines 195 to 203 in 6d3c5c0

	docker run -it --rm \
	--privileged \
	--network host \
	${DOCKER_ARGS[@]} \
	-v $ISAAC_ROS_DEV_DIR:/workspaces/isaac_ros-dev \
	-v /dev/*:/dev/* \
	-v /etc/localtime:/etc/localtime:ro \
	--name "$CONTAINER_NAME" \
	--runtime nvidia \

I noticed run_dev.sh's Docker container works (torch.cuda.is_available() returns True) if I replace --runtime nvidia with --gpus all. I also noticed in the dev environment setup guide (https://nvidia-isaac-ros.github.io/getting_started/dev_env_setup.html) that nvidia-container-runtime is deprecated. Is using --gpus all more suitable on newer versions of Docker?

[hemalshahNV]

--gpus all should enable the same runtime behavior but need to confirm with the nvidia-container-runtime engineers. Thanks for the heads up.

[YuminosukeSato]

Has this been changed?
If I use --gpus instead of --runtime in run_dev.sh, you will get an error like this.

docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: initialization error: load library failed: libnvidia-ml.so.1: cannot open shared object file: no such file or directory: unknown.

[jaiveersinghNV]

@Buddies-as-you-know , could you confirm what version of the CUDA Drivers you have installed? The missing libnvidia-ml.so.1 library should be included as part of a proper CUDA installation.