Windows+WSL2+GPU CUDA+Docker

参考：

1. CUDA on WSL User Guide
2. NVIDIA CUDA Installation Guide for Linux
3. WSL 2 GPU Support for Docker Desktop on NVIDIA GPUs - Docker

更新2022-04-24

不需要任何特殊的操作，就可以在Docker Desktop中使用GPU了

• 操作系统
  

• Docker Desktop
  

  • GPU驱动
    

• 满足上面的要求后，执行命令：docker run -it --gpus=all --rm nvidia/cuda:11.4.2-base-ubuntu20.04 nvidia-smi
  

1.操作系统

• 我采用的版本是windows10 教育版 2004(20262)。Ensure that you install Build version 20145 or higher.

• 系统必须安装有nvidia的卡，同时安装对应wsl2的驱动。

2.安装WSL2

参考：将WSL2作为生产力工具

3.安装Docker

在wsl2中：
curl https://get.docker.com | sh
关于Docker Desktop：
截止到2020.11.27，不支持Docker Desktop
参考：https://docs.nvidia.com/cuda/wsl-user-guide/index.html#known-limitations

4.安装cuda

从https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&Distribution=WSL-Ubuntu&target_version=2.0下载相应的安装程序

• 在WSL2中：
  注意：我使用的是Ubuntu18.04的镜像，如果是其他版本，请修改命令中的版本号

cd
apt-get install -y gcc g++ make
./cuda_11.1.0_455.23.05_linux.run

cd ./NVIDIA_CUDA-11.1_Samples/1_Utilities/deviceQuery
make
./deviceQuery

5.nvidia-docker toolkit

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
# To get access to experimental features such as CUDA on WSL or the new MIG capability on A100, you may want to add the experimental branch to the repository listing
curl -s -L https://nvidia.github.io/libnvidia-container/experimental/$distribution/libnvidia-container-experimental.list | sudo tee /etc/apt/sources.list.d/libnvidia-container-experimental.list

apt-get update
apt-get install -y nvidia-docker2
service docker stop
service docker start

6.验证

docker run --rm --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark

报错：

docker: Error response from daemon: cgroups: cannot find cgroup mount destination: unknown.

参考：https://blog.csdn.net/ren7wei/article/details/80855779

mkdir /sys/fs/cgroup/systemd
mount -t cgroup -o none,name=systemd cgroup /sys/fs/cgroup/systemd

再次运行验证命令，通过

安装脚本可以参考：https://blog.csdn.net/fleaxin/article/details/110310874