These instructions are from 2020 and slightly outdated now. They will be updated.

This is an example setup of a machine that should run GPU computing for BIIGLE. It is based on a clean install of Ubuntu 18.04 with the user ubuntu and an NVIDIA GPU.

  1. Run sudo apt update && sudo apt upgrade.

  2. Install the GPU drivers (adapted from the TensorFlow instructions):

    $ wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
$ sudo dpkg -i cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
$ rm cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
$ sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
$ sudo apt-get update
$ sudo apt-get install -y --no-install-recommends cuda-drivers

  3. Reboot the machine to activate the driver.

  4. Install Docker. If you are in a cloud environment, add the respository like this so it is not deleted when a new machine is booted:

    echo "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
    $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list

  5. Install nvidia-docker2 (Docker Compose does no support the newest version of nvidia-container-toolkit yet so we install the older version):

    $ 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

# install package maintainers version of /etc/docker/daemon.json
$ sudo apt-get update && sudo apt-get install -y --no-install-recommends nvidia-docker2

  6. Run sudo systemctl restart docker.

  7. Run sudo usermod -aG docker ubuntu to add the ubuntu user to the docker group.

  8. Install Docker Compose as a Docker container.

  9. Log out of the machine and back in.

Now you can follow the installation instructions to install BIIGLE on the machine.