WSL2 GPU Guide

Instructions for enabling GPU with Tensorflow on a WSL2 virtual machine.

Maintainer: @Avditvs (Independent)

Install the latest Windows build

You first need to install the latest build of Windows (version 20145 or higher). To do so, you'll have to subscribe to the Microsoft Windows Insider Program

Then, install the latest build from the Fast Ring by following the provided instructions.

Ubuntu 18.04 running is WSL2

First, install WSL2 by following the instructions provided by the Microsoft documentation.
After that you can download the Ubuntu 18.04 distribution from the Windows Store.

You should now be able to launch it from the Windows terminal or the Windows start menu.

Finally check that the kernel version is superior than 4.19.121 with the following command uname -r. If not, that means that the latest windows build is not installed.

Install the NVIDIA driver for WSL2

Using CUDA with WSL2 requires specific drivers. Install them by following these instructions.

Once installed, you can make sure that the driver is working and check the version by running the command in a Windows PowerShell terminal : nvidia-smi

Note : You may encounter issues the 465.12, if this version is not working, revert to a previous WSL2 enabled one (460.20 for instance).

Install the libraries

The last step is to install the libraries inside the Ubuntu 18.04 virtual machine. This can be done by running the commands below in the Ubuntu terminal.

# Add NVIDIA package repositories
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.1.243-1_amd64.deb
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo dpkg -i cuda-repo-ubuntu1804_10.1.243-1_amd64.deb
sudo apt-get update
wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt-get update

# Install development and runtime libraries (~4GB)
sudo apt-get install --no-install-recommends \
    cuda-10-1 \
    libcudnn7=7.6.5.32-1+cuda10.1  \
    libcudnn7-dev=7.6.5.32-1+cuda10.1


# Install TensorRT. Requires that libcudnn7 is installed above.
sudo apt-get install -y --no-install-recommends libnvinfer6=6.0.1-1+cuda10.1 \
    libnvinfer-dev=6.0.1-1+cuda10.1 \
    libnvinfer-plugin6=6.0.1-1+cuda10.1


#Add installation directories of libraries to this environment variable below.
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/cuda/include:/usr/local/cuda-10.2/targets/x86_64-linux/lib"

Check if everything is working

You can check if the GPU is available by running the following command inside a Python terminal (in Ubuntu):

import tensorflow as tf
gpus = tf.config.list_logical_devices(
    device_type='GPU'
)
print(gpus)