『heqingchun-ubuntu系统下安装cuda与cudnn』

本文链接：https://blog.csdn.net/heqingchun16/article/details/133787088

ubuntu系统下安装cuda与cudnn

一、安装依赖

1.更新

sudo apt update

sudo apt upgrade -y

2.基础工具

sudo apt install -y build-essential

二、安装CUDA

1.文件下载

网址

https://developer.nvidia.com/cuda-toolkit-archive

依次点击

(1)“CUDA Toolkit 11.6.2”

(2)“Linux”

(3)“x86_64”

(4)“Ubuntu”

(5)“20.04”

(6)“runfile(local)”

在"Installation Instructions:"下方为下载安装指令
下载指令(文件需下载到英文路径)，如：

cd /home/heqingchun/soft/nvidia

wget https://developer.download.nvidia.com/compute/cuda/11.6.2/local_installers/cuda_11.6.2_510.47.03_linux.run

在“/home/heqingchun/soft/nvidia”路径中下载得到“cuda_11.6.2_510.47.03_linux.run”文件
以下是安装时使用的指令

sudo sh cuda_11.6.2_510.47.03_linux.run

2.cuda安装

(1)赋予可执行权限

chmod 755 cuda_11.6.2_510.47.03_linux.run

(2)运行安装

sudo sh cuda_11.6.2_510.47.03_linux.run

期间会弹出对话框，需手动输入"accept"回车，在之后再弹出对话框中取消勾选“Driver”

CUDA Installer                                                               │
│ - [ ] Driver                                                                 │
│      [ ] 510.47.03                                                           │
│ + [X] CUDA Toolkit 11.6                                                      │
│   [X] CUDA Samples 11.6                                                      │
│   [X] CUDA Demo Suite 11.6                                                   │
│   [X] CUDA Documentation 11.6                                                │
│   Options                                                                    │
│   Install

向下选择"install"后等待安装完毕即可。
安装完毕信息：

===========
= Summary =
===========

Driver:   Not Selected
Toolkit:  Installed in /usr/local/cuda-11.6/

Please make sure that
 -   PATH includes /usr/local/cuda-11.6/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-11.6/lib64, or, add /usr/local/cuda-11.6/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-11.6/bin
***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least 510.00 is required for CUDA 11.6 functionality to work.
To install the driver using this installer, run the following command, replacing <CudaInstaller> with the name of this run file:
    sudo <CudaInstaller>.run --silent --driver

Logfile is /var/log/cuda-installer.log

(3)配置环境变量

str='export PATH=/usr/local/cuda-11.6/bin:"$"PATH' && \
sudo sh -c "echo $str >> /etc/profile" && \
source /etc/profile && \
str='export LD_LIBRARY_PATH=/usr/local/cuda-11.6/lib64:"$"LD_LIBRARY_PATH' && \
sudo sh -c "echo $str >> /etc/profile" && \
source /etc/profile

(4)重启电脑

3.验证安装

(1)版本信息

nvcc -V

显示如下:

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Tue_Mar__8_18:18:20_PST_2022
Cuda compilation tools, release 11.6, V11.6.124
Build cuda_11.6.r11.6/compiler.31057947_0

(2)库信息

cat /usr/local/cuda/version.json

显示如下:

{
   "cuda" : {
      "name" : "CUDA SDK",
      "version" : "11.6.20220318"
   },
   "cuda_cccl" : {
      "name" : "CUDA C++ Core Compute Libraries",
      "version" : "11.6.55"
   },
   "cuda_cudart" : {
      "name" : "CUDA Runtime (cudart)",
      "version" : "11.6.55"
   },
   "cuda_cuobjdump" : {
      "name" : "cuobjdump",
      "version" : "11.6.124"
   },
   "cuda_cupti" : {
      "name" : "CUPTI",
      "version" : "11.6.124"
   },
   "cuda_cuxxfilt" : {
      "name" : "CUDA cu++ filt",
      "version" : "11.6.124"
   },
   "cuda_demo_suite" : {
      "name" : "CUDA Demo Suite",
      "version" : "11.6.55"
   },
   "cuda_gdb" : {
      "name" : "CUDA GDB",
      "version" : "11.6.124"
   },
   "cuda_memcheck" : {
      "name" : "CUDA Memcheck",
      "version" : "11.6.124"
   },
   "cuda_nsight" : {
      "name" : "Nsight Eclipse Plugins",
      "version" : "11.6.124"
   },
   "cuda_nvcc" : {
      "name" : "CUDA NVCC",
      "version" : "11.6.124"
   },
   "cuda_nvdisasm" : {
      "name" : "CUDA nvdisasm",
      "version" : "11.6.124"
   },
   "cuda_nvml_dev" : {
      "name" : "CUDA NVML Headers",
      "version" : "11.6.55"
   },
   "cuda_nvprof" : {
      "name" : "CUDA nvprof",
      "version" : "11.6.124"
   },
   "cuda_nvprune" : {
      "name" : "CUDA nvprune",
      "version" : "11.6.124"
   },
   "cuda_nvrtc" : {
      "name" : "CUDA NVRTC",
      "version" : "11.6.124"
   },
   "cuda_nvtx" : {
      "name" : "CUDA NVTX",
      "version" : "11.6.124"
   },
   "cuda_nvvp" : {
      "name" : "CUDA NVVP",
      "version" : "11.6.124"
   },
   "cuda_samples" : {
      "name" : "CUDA Samples",
      "version" : "11.6.101"
   },
   "cuda_sanitizer_api" : {
      "name" : "CUDA Compute Sanitizer API",
      "version" : "11.6.124"
   },
   "libcublas" : {
      "name" : "CUDA cuBLAS",
      "version" : "11.9.2.110"
   },
   "libcufft" : {
      "name" : "CUDA cuFFT",
      "version" : "10.7.2.124"
   },
   "libcurand" : {
      "name" : "CUDA cuRAND",
      "version" : "10.2.9.124"
   },
   "libcusolver" : {
      "name" : "CUDA cuSOLVER",
      "version" : "11.3.4.124"
   },
   "libcusparse" : {
      "name" : "CUDA cuSPARSE",
      "version" : "11.7.2.124"
   },
   "libnpp" : {
      "name" : "CUDA NPP",
      "version" : "11.6.3.124"
   },
   "libnvjpeg" : {
      "name" : "CUDA nvJPEG",
      "version" : "11.6.2.124"
   },
   "nsight_compute" : {
      "name" : "Nsight Compute",
      "version" : "2022.1.1.2"
   },
   "nsight_systems" : {
      "name" : "Nsight Systems",
      "version" : "2021.5.2.53"
   },
   "nvidia_driver" : {
      "name" : "NVIDIA Linux Driver",
      "version" : "510.47.03"
   }
}

(3)计算能力

cd /usr/local/cuda/extras/demo_suite

./deviceQuery

显示:

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 11.6, CUDA Runtime Version = 11.6, NumDevs = 1, Device0 = NVIDIA GeForce RTX 3050 Laptop GPU
Result = PASS

CUDA安装完毕

三、安装cuDNN

1.文件下载

网址

https://developer.nvidia.com/rdp/cudnn-archive

依次点击

(1)“Download cuDNN v8.4.0 (April 1st, 2022), for CUDA 11.x”

(2)“Local Installer for Linux x86_64 (Tar)”

注:需要登陆，登陆成功后即可下载
下载得到“cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive.tar.xz”文件放入“/home/heqingchun/soft/nvidia”目录

2.cuDNN安装

进入文件所在目录、解压文件、解压后进入文件夹、拷贝文件

cd /home/heqingchun/soft/nvidia

tar -xvf cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive.tar.xz

cd cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive

sudo cp include/* /usr/local/cuda-11.6/include && \
sudo cp -P lib/* /usr/local/cuda-11.6/lib64 && \
sudo chmod a+r /usr/local/cuda-11.6/include/cudnn*.h /usr/local/cuda-11.6/lib64/libcudnn*

重启电脑

3.验证安装

cat /usr/local/cuda/include/cudnn_version.h

显示如下:

/**
 * \file: The master cuDNN version file.
 */

#ifndef CUDNN_VERSION_H_
#define CUDNN_VERSION_H_

#define CUDNN_MAJOR 8
#define CUDNN_MINOR 4
#define CUDNN_PATCHLEVEL 0

#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)

#endif /* CUDNN_VERSION_H */

cuDNN安装完毕
ubuntu系统下安装cuda与cudnn-完毕