【无标题】

导入ubuntu20.04镜像

sudo docker load -i ubuntu.tar

 在home/nvidia下新建文件夹

mkdir share_opt

sudo docker run -id --name=Apollo --net=host --runtime nvidia -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix -v /home/nvidia/share_opt:/share_opt -v /tmp/argus_socket:/tmp/argus_socket --device=/dev/video0 --device=/dev/video2 --device=/dev/video4 --device=/dev/video6 --privileged=true ubuntu:20.04_new

报错：

docker: Error response from daemon: unknown or invalid runtime name: nvidia.

和

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-container-runtime/$distribution/nvidia-container-runtime.list | sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list
sudo apt-get update

sudo apt-get install nvidia-container-runtime

报错：

 nvidia-container-toolkit : Depends: nvidia-container-toolkit-base (= 1.13.5-1) but it is not going to be installed

sudo apt --fix-broken install

选D： *** daemon.json (Y/I/N/O/D/Z) [default=N] ? D

sudo apt-get install nvidia-container-runtime

 nvidia-container-runtime 已经从3.9.0-1 升级成了3.13.0-1

暂时跳过 --runtime nvidia ，GPU加速，建立容器

sudo docker run -id --name=Apollo --net=host -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix -v /home/nvidia/share_opt:/share_opt -v /tmp/argus_socket:/tmp/argus_socket --device=/dev/video0 --device=/dev/video2 --device=/dev/video4 --device=/dev/video6 --privileged=true ubuntu:20.04_new

sudo apt-get update

sudo apt-get upgrade

sudo apt install software-properties-common

sudo apt-get update
sudo apt-add-repository multiverse
sudo apt-get update
sudo apt-get install nvidia-driver-455

export CUDA_HOME=/usr/local/cuda-11.4
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH
export PATH=$CUDA_HOME/bin:$PATH
export CPATH=$CUDA_HOME/include:$CPATH

在into.sh中配置挂载目录

-v /usr/local/cuda-11.4:/usr/local/cuda \

1. 安装 NVIDIA 容器工具包：

   • 确保已经安装了 NVIDIA 容器工具包（NVIDIA Container Toolkit），它是 Docker 与 NVIDIA GPU 驱动之间的重要桥梁，提供了 nvidia-container-runtime。

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-get update
 
sudo apt-get install -y nvidia-container-toolkit

1. 配置 Docker 以使用 NVIDIA 容器运行时：

   • 安装 NVIDIA 容器工具包后，需要配置 Docker 以使用 nvidia-container-runtime。通常，安装过程中会自动完成配置。您可以通过检查 /etc/docker/daemon.json 文件确认是否已正确配置。配置应类似如下：

     {
         "runtimes": {
             "nvidia": {
                 "path": "/usr/bin/nvidia-container-runtime",
                 "runtimeArgs": []
             }
         }
     }

     sudo systemctl restart docker

现在可以使用docker run --runtime=nvidia -it your_gpu_image

docker run -it --env PATH="/usr/local/cuda-11.4/bin:$PATH" your_image_name bash

export PATH="/usr/local/cuda-11.4/bin:$PATH"

挂载权限？

ls -ld /usr/local/cuda-11.4

查看GPU版本

cat /proc/driver/nvidia/version

 sudo docker run --runtime=nvidia -v /usr/local/cuda-11.4:/usr/local/cuda-11.4 -it --name=apollo5 apollo_docker:4.19