Features Cuda compilation tools, release 10.0, V10.0.130
NVIDIA SMI Driver version 410.48
NOTE: This quickstart guide assumes you have installed docker-ce, NVIDIA driver, cuda, nvidia-cuda-toolkit and nvidia-docker.
- Clone the repository
git clone git@github.com:JRWu/gpu-compute-environment.git
- Change directories into the repository.
cd gpu-compute-environment/
- Rename the project from 'project:' to 'my_project:'. You can set 'my_project' to be whatever you want. This step is important if multiple users on a single system wish to use this gpu-compute-environment.
sed -i -e 's#project:#my_project:#g' docker-compose.yml
- Build the container.
docker-compose up --build -d
- Shell into the gpu compute environment using the <my_project> name you set before.
docker-compose exec my_project bash
NOTE: This guide assumes NVIDIA hardware running ontop of Ubuntu 18.04.
- Install stable docker-ce.
sudo apt-get update
sudo apt-get install -y apt-transport-https ca-certificates curl 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-get update
sudo apt-get install docker-ce
(Optional) Perform post-install docker instructions in order to run docker as root.
sudo groupadd docker
sudo usermod -aG docker $USER
NOTE: You must restart your shell or your computer for the docker permission changes to take effect.
- Download and Install the NVIDIA driver from: https://www.nvidia.com/content/DriverDownload-March2009/confirmation.php?url=/XFree86/Linux-x86_64/410.66/NVIDIA-Linux-x86_64-410.66.run&lang=us&type=TITAN
# Assuming you downloaded into the Downloads directory:
cd /home/$USER/Downloads
chmod +x NVIDIA-Linux-x86_64-410.66.run
sudo ./NVIDIA-Linux-x86_64-410.66.run
# Follow the recommended instructions for the installation.
- Download and Install the NVIDIA cuda toolkit from: https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=deblocal
Assuming you downloaded into the Downloads directory:
cd /home/$USER/Downloads
sudo dpkg -i cuda-repo-ubuntu1804-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
sudo dpkg -i cuda-repo-ubuntu1804-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
sudo apt-get update
sudo apt install -y cuda nvidia-cuda-toolkit
At this point, you must reboot your PC in order for the driver changes to take effect.
- Install the nvidia-docker runtime.
# If you have nvidia-docker 1.0 installed: we need to remove it and all existing GPU containers
docker volume ls -q -f driver=nvidia-docker | xargs -r -I{} -n1 docker ps -q -a -f volume={} | xargs -r docker rm -f
sudo apt-get purge -y nvidia-docker
# Add the package repositories
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
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-get update
# Install nvidia-docker2 and reload the Docker daemon configuration
sudo apt-get install -y nvidia-docker2
sudo pkill -SIGHUP dockerd
# Test nvidia-smi with the latest official CUDA image
docker run --runtime=nvidia --rm nvidia/cuda:9.0-base nvidia-smi
If the command did not throw an error, you may proceed to the Quickstart portion.
Sometimes the command:
ldconfig
Must be run within the container before the GPU is usable.
Add notes for how to log into nvcr.
Add notes for how to debug containers.
To delete ALL Docker-related volumes, images, networks and containers run the following script. WARNING: THIS STEP CANNOT BE REVERSED.
bash tools/clean_docker.sh