Nvidia Docker安装与测试

安装docker ce (ubuntu可以直接安装docker.io)

$ sudo apt install docker.io

安装nvidia driver

$ cat /etc/issue
$ sudo lshw -C display

$ sudo apt update
$ sudo apt search nvidia-driver

$ nvidia-detector
$ sudo apt install nvidia-driver-440
$ reboot

安装nvidia docker

$ curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
$ curl -s -L https://nvidia.github.io/nvidia-docker/ubuntu18.04/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

$ sudo apt update
$ sudo apt-get install nvidia-docker2
$ sudo pkill -SIGHUP dockerd

# 如果报错，先执行nvidia-smi看一下host的cuda version，再对应跑一下
$ docker run --runtime=nvidia --rm nvidia/cuda:10.2-devel nvidia-smi

• 参考: https://blog.csdn.net/BigData_Mining/article/details/99681168

修改Daemon.json

• 修改docker配置（如果采用Swarm启动nvidia docker）

$ cat /etc/docker/daemon.json 
{
    "runtimes": {
        "nvidia": {
            "path": "nvidia-container-runtime",
            "runtimeArgs": []
        }
    },
    "default-runtime": "nvidia"
}

• 如果修改了上述配置，则执行下面命令，加载OS参数

$ sudo systemctl daemon-reload
$ sudo systemctl restart docker