4 steps to get a docker running with GPU on Ubuntu 20.04

0-4-steps-get-docker-with-gpu-on-ubuntu-2004.md

⚠️ The following instruction is written on 25 Jul 2020 (= old).

4 steps to get a docker running with GPU on Ubuntu 20.04

Environment

• Ubuntu 20.04
• CUDA 10.2
• Docker Latest

Four Steps

1. 1-install-first.sh
2. 2-install-cuda.sh -> Reboot
3. 3-install-docker.sh -> Reboot
4. 4-install-nvidia.sh

1-install-first.sh

#!/bin/sh

sudo apt update && sudo apt -y upgrade
sudo apt install -y build-essential

2-install-cuda.sh

#!/bin/sh

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-10-2-local-10.2.89-440.33.01/7fa2af80.pub
sudo apt update
sudo apt -y install cuda-drivers
sudo systemctl set-default  multi-user.target
sudo reboot

3-install-docker.sh

#!/bin/sh

sudo apt -y install \
    apt-transport-https \
    ca-certificates \
    curl \
    gnupg-agent \
    software-properties-common
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
sudo add-apt-repository \
    "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
    $(lsb_release -cs) \
    stable"
sudo apt update
sudo apt -y install docker-ce docker-ce-cli containerd.io
sudo usermod -aG docker $USER
sudo reboot

4-install-nvidia.sh

#!/bin/sh

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
sudo apt update && sudo apt -y install nvidia-container-toolkit
sudo systemctl restart docker
docker run --gpus all --pid host nvidia/cuda:10.2-runtime nvidia-smi

Why is this so complicated?
In step 4, altough I have ubuntu 20.04, the .list file contains references to ubuntu 18. Later on the apt update fails..
Any idea why?

@eyalhir74 if you ran all the .sh files verbatim you most likely wouldn't complete the update correctly. From my experience alot of getting it to work properly is knowing what nvidia drivers (and thusly cuda version) you are running. Check that by simply running $ nvidia-smi and checking the first line of the output.
I believe that as of driver version 430+ cuda is built in, so step 2 above is not required. (That also assumes you have the appropriate nvidia drivers installed prior to tackling the docker install)

Go figure, I found this article from IBM (of all places) very straight forward: Installing nvidia-docker2 - follow x86 Ubuntu instructions
Then subbing in the cuda version (to do the verification) that matches your output from before (with the assist from: Supported CUDA tags)

@eyalhir74 I followed these steps that redirected me here.

• Install a correct NVIDIA driver on your host. Reboot.
• Go here and install the correct packages depending on your system OR replace the wrong Ubuntu version in the link with the correct one and it should work. I think you do not need CUDA on your host system for the docker to work but I might be wrong

I suggest you to reboot between the steps and verify your installations. If it does not work, just find another guide by googling "nvidia docker ubuntu 20.04" :)

I'm on Ubuntu 22.04

I believe that as of driver version 430+ cuda is built in, so step 2 above is not required

Because of this comment, I started at step 3 and it worked out