# check if GPU is recognizable within container
docker run --rm -it --gpus=all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark
CUDA on WSL :: CUDA Toolkit Documentation
Lots of DS folks use Windows for some reason. Also lots of people buy gaming desktops to play with CUDA-enabled Deep Learning. Couple that with official support of such machines by Microsoft and OEMs it makes it lucrative to build support for our infrastructure stack for docker on windows.
Announcing CUDA on Windows Subsystem for Linux 2 | NVIDIA Developer Blog
https://blogs.windows.com/windowsdeveloper/2020/06/17/gpu-accelerated-ml-training-inside-the-windows-subsystem-for-linux/
linux support on Win10 is much better than before (check WSL) but it still no smooth experience
nvidia-runtime for docker on windows is still in Beta
no support for Docker on WSL 2 yet
https://docs.nvidia.com/cuda/wsl-user-guide/index.html#known-limitations