This happened when running docker commands on my ASUS version of the Spark. I suspect it is specific to ASUS. It happened on both of my ASUS machines, out of the box. Once it was cleaned up it works fine. I will soon do a fresh install from their recovery iso, and will see if the bad file is put down from that.
Has anybody managed to run a container with GPU support? I tried this:
docker run -d --restart no \ --gpus=all \ -p 11434:11434 \ --add-host=host.docker.internal:host-gateway \ -v ollama:/root/.ollama \ --name ollama \ ollama/ollama:latest
But I get this error:
time=2025-11-07T20:43:04.384Z level=INFO source=runner.go:67 msg=“discovering available GPUs…” time=2025-11-07T20:43:04.384Z level=INFO source=server.go:400 msg=“starting runner” cmd=“/usr/bin/ollama runner --ollama-engine --port 34271” time=2025-11-07T20:43:34.413Z level=INFO source=runner.go:442 msg=“failure during GPU discovery” OLLAMA_LIBRARY_PATH=“[/usr/lib/ollama /usr/lib/ollama/cuda_jetpack5]” extra_envs=map error=“failed to finish discovery before timeout”
I’ve even ripped out all Docker-related packages and installed a brand new Docker Engine from here Ubuntu | Docker Docs and I get the same result.
nvidia-container-toolkit is installed.
EDIT: Nevermind, this might be an Ollama issue. I’ve added --runtime=nvidia to the container options, and it seems to run just fine. This command also works well:
docker run -it --runtime=nvidia --gpus=all nvcr.io/nvidia/cuda:13.0.1-devel-ubuntu24.04 nvidia-smi