docker+nvidia

解决docker容器无法驱动GPU问题：

1. 检查本机GPU是否存在问题

nvidia-smi

检查结果

Wed Jul 10 17:46:53 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.01             Driver Version: 535.183.01   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA GeForce RTX 4060 Ti     Off | 00000000:01:00.0  On |                  N/A |
|  0%   53C    P8               2W / 160W |    536MiB /  8188MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|    0   N/A  N/A      1925      G   /usr/lib/xorg/Xorg                          238MiB |
|    0   N/A  N/A      2119      G   /usr/bin/gnome-shell                         30MiB |
|    0   N/A  N/A      2824      G   ...99,262144 --variations-seed-version      197MiB |
|    0   N/A  N/A      5932      G   ...erProcess --variations-seed-version       60MiB |
+---------------------------------------------------------------------------------------+

出现上述结果说明无误

2. 安装NVIDIA Container Toolkit 软件包

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

3. 配置docker的显卡支持属性

sudo nvidia-ctk runtime configure --runtime=docker

检查配置是否成功

{
    "registry-mirrors": [
        "https://xxxx.mirror.aliyuncs.com"
    ],
    "runtimes": {
        "nvidia": {
            "args": [],
            "path": "nvidia-container-runtime"
        }
    }
}

该文件是处理docker image加速时创建的，一般采用阿里云加速，加速方法见

阿里云-docker 加速

4.docker容器创建

总体

sudo docker run  -dit \
--gpus all \
-e NVIDIA_DRIVER_CAPABILITIES=all \
--name=melodic_docker2 \
-v /home/yearner:/home/yearner \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-v /dev/dri:/dev/dri \
--device=/dev/snd \
--device=/dev/dri/renderD128 \
-e DISPLAY=unix$DISPLAY \
-w /home/yearner fishros2/ros:melodic-desktop-full

其中

--gpus all \
-e NVIDIA_DRIVER_CAPABILITIES=all

是让容器支持GPU

-v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=unix$DISPLAY --device=/dev/dri/renderD128 -v /dev/dri:/dev/dri 

这里是驱动docker容器调用本地xhost接口显示