#如何在linux系统中安装英伟达驱动并配置CUDA和cuDNN

1. 安装驱动

   • 安装依赖

     sudo apt install build-essential libglvnd-dev pkg-config
     

   • 切换到命令行模式（服务器版本系统不需要）

     sudo telinit 3
     

   • 下载驱动并赋予执行权限，并安装

     wget URL
     sudo chmod +x NVIDIA-Linux-x86_64-xxx.xx.run
     sudo ./NVIDIA-Linux-x86_64-xxx.xx.run
     

     如果有报错则执行

     sudo apt install linux-headers-$(uname -r)
     #pve 系统改成pve开头
     #pve-headers-$(uname -r)
     

     安装完运行
     nvidia-smi
     有输出则成功

2. 安装CUDA
   下载CUDA并执行（我选择的11.8版本）

   wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run
   chmod +x cuda_11.8.0_520.61.05_linux.run
   sudo ./cuda_11.8.0_520.61.05_linux.run
   

3. 添加环境变量
   默认安装在/usr/local下/usr/local/cuda-11.8

   nano ~/.bashrc
   ##############
   #末尾添加
   export PATH=/usr/local/cuda-11.8/bin:$PATH
   export LD_LIBRARY_PATH=/usr/local/cuda-11.8/lib64:$LD_LIBRARY_PATH
   

   软连/usr/local/cuda

   sudo ln -s /usr/local/cuda-11.8 /usr/local/cuda
   

   验证安装
   nvcc -V
   应该输出：

   nvcc: NVIDIA (R) Cuda compiler driver
   Copyright (c) 2005-2022 NVIDIA Corporation
   Built on Wed_Sep_21_10:33:58_PDT_2022
   Cuda compilation tools, release 11.8, V11.8.89
   Build cuda_11.8.r11.8/compiler.31833905_0
   

4. 安装cuDNN
   cuDNN下载地址

tar -xvf cudnn-linux-x86_64-8.6.0.163_cuda11-archive.tar.xz
cd cudnn-linux-x86_64-8.6.0.163_cuda11-archive
sudo cp include/cudnn*    /usr/local/cuda/include 
sudo cp lib/libcudnn*    /usr/local/cuda/lib64 
sudo chmod a+r /usr/local/cuda/include/cudnn*   /usr/local/cuda/lib64/libcudnn*

cuDNN安装完成，查看输出

cat /usr/local/cuda/include/cudnn_version.h | grep CUDNN_MAJOR -A 2

应该显示

#define CUDNN_MAJOR 8
#define CUDNN_MINOR 9
#define CUDNN_PATCHLEVEL 7
--
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)

至此，全部安装完成