install wsl2

prerequisites

Windows 10 version 2004 and later (build 19041 and later) or Windows 11

通过按 Windows 徽标键 + R ---输入winver，单击确定，检查你的 Windows 版本

Please upgrade if the internal version is lower than required . Log in to your windows account and upgrade windows to the preview version, (the process may take 1-2 hours)

After the upgrade is complete, verify whether the internal version is lower than 19041. If it is still lower , it means that you have selected the wrong upgrade channel. Be sure to choose the Dev channel, the latest channel, and get the latest version

Install preview nvdia driver

GPU in Windows Subsystem for Linux (WSL) | NVIDIA Developer

You need to log in to your nvidia account to download. After logging in to your account, download the nvidia version that suits you.

Enable windows subsystem features

Install WSL2

1. Administrator PowerShell

dism.exe /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart

Restart the computer

2. Admin PowerShell sets WSL2 as the default version

wsl --set-default-version 2

Prompt "WSL 2 needs to update its kernel components". Download and install the WSL2 Llinx kernel

3. Then enter the Microsoft store to download and install the corresponding ubuntu version. Or go here to choose a specific Linux version

Two ways to log in to wsl2

If the system cannot enter and reports an error , you need to download and install the WSL2 Llinx kernel

Make sure the linux kernel of WSL2 is 4.19.121+

uname -a

verify-wsl2

↓↓↓↓↓↓↓↓↓↓↓ Windows Administrator PowerShell

wsl --list --verbose

As long as this version is 2, the wsl2 installation is successful. (Only 2 can call the GPU)

install docker

The following code is executed line by line

sudo apt-get update

sudo apt-get install \
    apt-transport-https \
    ca-certificates \
    curl \
    gnupg \
    lsb-release

curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg

echo \
  "deb [arch=amd64 signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu \
  $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null

sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io

Install nvidia-docker

Use nvdia-docker directly, the current system does not need to install cuda

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/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
curl -s -L https://nvidia.github.io/libnvidia-container/experimental/$distribution/libnvidia-container-experimental.list | sudo tee /etc/apt/sources.list.d/libnvidia-container-experimental.list

sudo apt-get update
sudo apt-get install -y nvidia-docker2

restart docker

sudo service docker stop
sudo service docker start

Verify that nvidia-docekr can call GPU

Create a tensorflow container

sudo docker run --runtime=nvidia  --rm -it --name tensorflow-1.14.0 tensorflow/tensorflow:1.14.0-gpu-py3

If an error is reported, it is because the latest version of nvidia-docker has a BUG (2022-2-10 re-test this bug has been resolved)

docker: Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: Running hook #1:: error running hook: exit status 1, stdout: , stderr: nvidia-container-cli: initialization error: driver error: failed to process request: unknown.

Downgrade some packages

sudo apt-get install nvidia-docker2:amd64=2.5.0-1 \
           libnvidia-container-tools:amd64=1.3.3-1 \
           nvidia-container-runtime:amd64=3.4.2-1 \
           libnvidia-container1:amd64=1.3.3-1 \
           nvidia-container-toolkit:amd64=1.4.2-1

run the container again

sudo docker run --runtime=nvidia  --rm -it --name tensorflow-1.14.0 tensorflow/tensorflow:1.14.0-gpu-py3

Test if the GPU is available

python
import tensorflow as tf
print(tf.test.is_gpu_available())

Windows wsl operation skills and commands

Windows wsl operation skills and commands_SUNbrightness' Blog-CSDN Blog_wsl Skills

reference link

Install WSL | Microsoft Docs

CUDA on WSL :: CUDA Toolkit Documentation

nvidia-docker 2.6.0-1 - not working on Ubuntu WSL2 · Issue #1496 · NVIDIA/nvidia-docker · GitHub

Install Docker Engine on Ubuntu | Docker Documentation

The ubuntu subsystem of Windows system WSL2 installs docker, nvidia-docker invokes GPU