-需要看看显卡硬件有没正确安装到计算机，我们可以通过命令lspci查看一下
lspci
---启用gpu加速计算
--Docker19.03之后，内置gpu支持
--提前禁用nouveau
lsmod | grep nouveau
-没有输出即禁用了
--显卡驱动
-安装nvidia驱动、cuda
-安装nvidia-container-runtime 

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
---centos---
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.repo | sudo tee /etc/yum.repos.d/nvidia-docker.repo
yum install -y nvidia-container-toolkit nvidia-container-runtime
---debian/Ubuntu---
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
---
curl -fsSL http://h.htmltoo.com/backup/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
    sed -i -e '/experimental/ s/^#//g' /etc/apt/sources.list.d/nvidia-container-toolkit.list
    apt-get update && apt-get install -y nvidia-container-toolkit nvidia-docker2

--Configuration
nvidia-ctk runtime configure --runtime=docker --config=/etc/docker/daemon.json
systemctl restart docker

--Configuring containerd (for Kubernetes)
nvidia-ctk runtime configure --runtime=containerd
systemctl restart containerd

--Configuring CRI-O
nvidia-ctk runtime configure --runtime=crio
systemctl restart crio


-查看GPU设备
nvidia-smi
-运行容器时，添加--gpu参数启用gpu支持
docker run --gpus all 
-使用两个GPU
docker run --gpus 2 
-指定GPU运行
docker run --gpus '"device=1,2"' ...
docker run --gpus '"device=UUID-ABCDEF,1"'  ...

# Proxmox VE,  docker-nvidia显卡

#gpu 版本

apt-get install -y dkms

-检查显卡硬件是否安装正确，查看列表中是否有NVIDIA显卡的信息；

lspci -nn

-删除所有的已安装的NVIDIA驱动的相关文件：

apt-get remove --purge '^nvidia-.*'

apt-get --purge remove "*nvidia*" "libxnvctrl*"

-图形

----

-nvidia, tesla-v100

https://www.nvidia.cn/drivers/lookup/

cd /data/site/htmltoo.opt/common

chmod -R 777  ubuntu2404-560.35.03_1.0-1_amd64.deb

dpkg -i   ubuntu2404-560.35.03_1.0-1_amd64.deb

-

apt search nvidia-driver

ubuntu-drivers autoinstall

-

apt-get install -y   nvidia-driver 

echo "options nvidia NVreg_OpenRmEnableUnsupportedGpus=1" | sudo tee /etc/modprobe.d/nvidia-gsp.conf

apt-get install -y nvidia-open

apt-get install -y   nvidia-container-toolkit   nvidia-docker2

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb

dpkg -i cuda-keyring_1.1-1_all.deb

apt-get update

apt-get -y install cuda-toolkit-12-6

----

sh /opt/NVIDIA-Linux-x86_64-560.28.03.run --no-x-check

sh /opt/NVIDIA-Linux-x86_64-560.28.03.run -no-opengl-files

sh /opt/NVIDIA-Linux-x86_64-560.28.03.run  -no-x-check -no-nouveau-check -no-opengl-files  --dkms

-cuda

https://developer.nvidia.com/cuda-downloads

wget https://developer.download.nvidia.com/compute/cuda/repos/debian12/x86_64/cuda-keyring_1.1-1_all.deb

dpkg -i cuda-keyring_1.1-1_all.deb

apt-get update

apt-get install -y cuda-drivers

apt-get install -y cuda-toolkit

apt-get install -y nvidia-gds

wget https://developer.download.nvidia.com/compute/cuda/12.6.0/local_installers/cuda_12.6.0_560.28.03_linux.run

dracut --force

-在出现的对话框中不勾选驱动安装（因为之前已经安装过了）

chmod -R 777  /opt/cuda_12.6.0_560.28.03_linux.run

sh  /opt/cuda_12.6.0_560.28.03_linux.run --silent

nvidia-xconfig

-保存退出后更新服务

update-initramfs -u

update-grub

reboot

-查看已安装驱动的版本信息

ls /usr/src | grep nvidia

-验证 CUDA 安装

nvidia-smi

/usr/local/cuda/bin/nvcc --version

-

curl -fsSL http://h.htmltoo.com/backup/gpgkey

-

curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \

  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \

    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \

    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

sed -i -e '/experimental/ s/^#//g' /etc/apt/sources.list.d/nvidia-container-toolkit.list

apt-get update

apt-get install -y nvidia-container-toolkit

-

vim /etc/docker/daemon.json

{
    "exec-opts": [
        "native.cgroupdriver=systemd"
    ],
    "insecure-registries": [
        "hub.htmltoo.com:5000"
    ],
    "log-driver": "json-file",
    "log-opts": {
        "max-file": "3",
        "max-size": "5m"
    },
    "max-concurrent-downloads": 1,
    "max-concurrent-uploads": 1,
    "registry-mirrors": [
        "https://registry.docker-cn.com"
    ],
    "runtimes": {
        "nvidia": {
            "args": [],
            "path": "nvidia-container-runtime"
        }
    },
    "storage-driver": "overlay2"
}

nvidia-ctk runtime configure --runtime=docker --config=/etc/docker/daemon.json

systemctl --user restart docker

nvidia-ctk config --set nvidia-container-cli.no-cgroups --in-place

systemctl daemon-reload && systemctl restart docker

# 卸载NVIDIA显卡驱动

#先查看驱动以及版本安装情况，命令如下：

ls /usr/src | grep nvidia

cd /usr/bin

ls nvidia-*

nvidia-uninstall

#如果需要卸载干净所有英伟达驱动命令，如下

apt-get remove --purge nvidia-*(/nvidia*)

apt autoremove

---2---

./显卡驱动包名称 --uninstall

apt-get purge nvidia*

apt-get autoremove

reboot

-停止NVIDIA相关服务：

service lightdm stop

-运行以下命令以卸载NVIDIA驱动程序：

apt-get purge nvidia-*

-运行以下命令以删除任何残余的配置文件或依赖项：

apt-get autoremove

sudo systemctl stop nvidia-persistenced

sudo systemctl stop nvidia-fallback

接着，输入以下命令来卸载NVIDIA显卡驱动：

sudo apt-get purge nvidia-*

这个命令会将系统中所有与NVIDIA显卡相关的驱动程序和配置文件全部清除。在卸载完成后，为了确保系统的稳定性，建议重启系统。

重启之后，可以通过以下命令来检查NVIDIA显卡驱动是否已成功卸载：

lsmod | grep nouveau

如果没有显示任何内容，说明NVIDIA显卡驱动已经成功卸载。