navidia环境安装

nvidia驱动安装

yum -y install kernel-devel epel-release dkms gcc kmod
lspci
lspci -v -s 00:06.0
http://pci-ids.ucw.cz/mods/PC/10de?action=help?help=pci 输入查询的网卡编号查看显卡版本

rpm -Uvh http://www.elrepo.org/elrepo-release-7.0-2.el7.elrepo.noarch.rpm

yum -y install kmod-nvidia-470xx.x86_64

rpm -ql kerner-devel
sh NVIDIA-Linux-x86_64-450.203.03.run --kernel-source-path=/usr/src/kernels/3.10.0-957.el7.x86_64/ --no-drm
rpm -qa | grep -E "kernel-devel|kernel-headers"
kernel-devel已经给出，一般是在/usr/src/kernels/目录下

nvidia-smi

cuda、cudnn

yum -y install gcc gcc-c++ kernel-devel

sudo sh cuda_11.0.2_450.51.05_linux.run

vim ~/.bashrc
#按o进入编辑模式，并在最后一行添加如下代码
export PATH=$PATH:/usr/local/cuda-11.0/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-11.0/lib64
export LIBRARY_PATH=$LIBRARY_PATH:/usr/local/cuda-11.0/lib64
source ~/.bashrc
nvcc --version

cuDNN Archive | NVIDIA Developer

tar -zxvf cudnn-11.0-linux-x64-v8.0.5.39.tgz
cp cuda/include/cudnn.h /usr/local/cuda/include
cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
 chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn* 
rpm -ivh libcudnn8-8.0.5.39-1.cuda11.0.x86_64.rpm 
rpm -ivh libcudnn8-devel-8.0.5.39-1.cuda11.0.x86_64.rpm
rpm -ivh libcudnn8-samples-8.0.5.39-1.cuda11.0.x86_64.rpm

nvidia-docker

yum install --downloadonly nvidia-docker2 --downloaddir=/tmp/nvidia


rpm -ivh libnvidia-container1-1.1.1-1.x86_64.rpm libnvidia-container-tools-1.1.1-1.x86_64.rpm
rpm -ivh nvidia-container-runtime-3.2.0-1.x86_64.rpm nvidia-container-toolkit-1.1.2-2.x86_64.rpm nvidia-docker2-2.3.0-1.noarch.rpm 
nvida-docker version

runtime配置

# 编辑daemon.json文件，并重启docker
vi /etc/docker/daemon.json
···
{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "/usr/bin/nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}
···

systemctl daemon-reload
systemctl restart docker