SiriBlog

siriyang的个人博客


  • 首页

  • 排行榜

  • 标签115

  • 分类37

  • 归档320

  • 关于

  • 搜索

Ubuntu下安装NVIDIA驱动及CUDA机器学习环境

发表于 2020-09-28 更新于 2021-10-29 分类于 计算机 , 环境 阅读次数: Valine:
本文字数: 7.2k 阅读时长 ≈ 7 分钟

安装NVIDIA驱动

查看可安装的硬件驱动:

1
ubuntu-drivers devices

== /sys/devices/pci0000:2c/0000:2c:00.0/0000:2d:00.0 ==
modalias : pci:v000010DEd00001BB1sv0000103Csd000011A3bc03sc00i00
vendor : NVIDIA Corporation
model : GP104GL [Quadro P4000]
manual_install: True
driver : nvidia-driver-418-server - distro non-free
driver : nvidia-driver-450-server - distro non-free recommended
driver : nvidia-driver-440-server - distro non-free
driver : nvidia-driver-390 - distro non-free
driver : nvidia-driver-450 - distro non-free
driver : nvidia-driver-435 - distro non-free
driver : xserver-xorg-video-nouveau - distro free builtin

安装驱动:

1
sudo ubuntu-drivers autoinstall

安装完成以后进行重启:

1
sudo reboot

查看驱动版本:

1
sudo dpkg --list | grep nvidia-*
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
ii  libnvidia-cfg1-450-server:amd64            450.51.06-0ubuntu0.18.04.2                   amd64        NVIDIA binary OpenGL/GLX configuration library
ii libnvidia-common-450-server 450.51.06-0ubuntu0.18.04.2 all Shared files used by the NVIDIA libraries
ii libnvidia-compute-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA libcompute package
ii libnvidia-compute-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVIDIA libcompute package
ii libnvidia-decode-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA Video Decoding runtime libraries
ii libnvidia-decode-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVIDIA Video Decoding runtime libraries
ii libnvidia-encode-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVENC Video Encoding runtime library
ii libnvidia-encode-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVENC Video Encoding runtime library
ii libnvidia-extra-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 Extra libraries for the NVIDIA Server Driver
ii libnvidia-fbc1-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA OpenGL-based Framebuffer Capture runtime library
ii libnvidia-fbc1-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVIDIA OpenGL-based Framebuffer Capture runtime library
ii libnvidia-gl-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA OpenGL/GLX/EGL/GLES GLVND libraries and Vulkan ICD
ii libnvidia-gl-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVIDIA OpenGL/GLX/EGL/GLES GLVND libraries and Vulkan ICD
ii libnvidia-ifr1-450-server:amd64 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA OpenGL-based Inband Frame Readback runtime library
ii libnvidia-ifr1-450-server:i386 450.51.06-0ubuntu0.18.04.2 i386 NVIDIA OpenGL-based Inband Frame Readback runtime library
ii nvidia-compute-utils-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA compute utilities
ii nvidia-dkms-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA DKMS package
ii nvidia-driver-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA Server Driver metapackage
ii nvidia-kernel-common-450-server 450.51.06-0ubuntu0.18.04.2 amd64 Shared files used with the kernel module
ii nvidia-kernel-source-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA kernel source package
ii nvidia-prime 0.8.8.2 all Tools to enable NVIDIA's Prime
ii nvidia-settings 440.82-0ubuntu0.18.04.1 amd64 Tool for configuring the NVIDIA graphics driver
ii nvidia-utils-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA Server Driver support binaries
ii xserver-xorg-video-nvidia-450-server 450.51.06-0ubuntu0.18.04.2 amd64 NVIDIA binary Xorg driver

参考资料:

  • 【nvidia】1.命令行方式安装nvidia显卡驱动

安装CUDA

下载CUDA
  前往下载页面选择和本机硬件系统相对应的CUDA版本,选好后会自动生成下载命令:

我的配置:Linux、x86_64、Ubuntu、18.04、runfile(local)

1
2
wget https://developer.download.nvidia.com/compute/cuda/11.1.0/local_installers/cuda_11.1.0_455.23.05_linux.run
sudo sh cuda_11.1.0_455.23.05_linux.run

安装CUDA
  由于我已经安装了驱动,在安装CUDA的时候会有一个提示,选择continue。
  之后会让你阅读声明,直接输入accept回车就行了。
  接下来会进入安装选项界面,可以配置一些安装参数。首先先把驱动取消勾选,其他保持默认即可。

  安装完成后显示如下:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
===========
= Summary =
===========

Driver: Not Selected
Toolkit: Installed in /usr/local/cuda-11.1/
Samples: Installed in /home/ubuntu/, but missing recommended libraries

Please make sure that
- PATH includes /usr/local/cuda-11.1/bin
- LD_LIBRARY_PATH includes /usr/local/cuda-11.1/lib64, or, add /usr/local/cuda-11.1/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-11.1/bin
***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least .00 is required for CUDA 11.1 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

验证CUDA

先配置环境变量:

1
2
export PATH="/usr/local/cuda-11.1/bin:$PATH" 
export LD_LIBRARY_PATH="/usr/local/cuda-11.1/lib64:$LD_LIBRARY_PATH"

然后输入:

1
nvcc -V

之前的配置正确的话会输出版本信息:

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Nov__3_21:07:56_CDT_2017
Cuda compilation tools, release 9.1, V9.1.85

编译samples例子:

1
2
3
4
5
6
7
8
9
#编译并测试设备 deviceQuery:
cd /usr/local/cuda-11.1/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery

#编译并测试带宽 bandwidthTest:
cd ../bandwidthTest
sudo make
./bandwidthTest

正常情况下会输出Result = PASS

将cuda的bin和lib写入系统环境:

打开~.bashrc文件在末尾追加:

1
2
3
export CUDA_HOME=/usr/local/cuda-11.1
export LD_LIBRARY_PATH=/usr/local/cuda-11.1/lib64:$LD_LIBRARY_PATH
export PATH=/usr/local/cuda-11.1/bin:$PATH

参考资料:

  • Linux安装CUDA的正确姿势

安装cuDNN

下载cuDNN

  下载地址:https://developer.nvidia.com/rdp/cudnn-archive

  选择对应CUDA版本的cuDNN:cuDNN Library for Linux (x86_64)

  经过多次尝试,官网的下载链接老是断开连接下载失败,于是在网上找到了一个镜像下载地址。只需要把下载链接中的版本号改为自己需要的就行了,然后复制到迅雷中下载,速度很快。下载好以后再传到服务器。

1
http://file.ppwwyyxx.com/nvidia/cudnn-11.0-linux-x64-v8.0.2.39.tgz

安装cuDNN

  如果是在官网下载,下载下来的文件格式为.solitairetheme8,要解压的话需要转换为.tgz格式,直接修改后缀即可。
  首先将压缩包解压:

1
tar -xvf cudnn-11.0-linux-x64-v8.0.2.39.tgz

  解压好以后会看到一个cuda文件夹,接下来执行如下命令:

1
2
3
4
5
6
7
# 复制cudnn头文件
sudo cp cuda/include/* /usr/local/cuda-11.1/include/
# 复制cudnn的库
sudo cp cuda/lib64/* /usr/local/cuda-11.1/lib64/
# 添加可执行权限
sudo chmod +x /usr/local/cuda-11.1/include/cudnn.h
sudo chmod +x /usr/local/cuda-11.1/lib64/libcudnn*

  整个安装过程实际上就是把cuDNN的头文件和静态库文件拷贝到CUDA的相应目录下。

验证cuDNN

1
cat /usr/local/cuda/include/cudnn.h

参考资料:

  • Ubuntu16.04下安装cuda和cudnn的三种方法(亲测全部有效)
  • ubuntu安装cudnn
  • Ubuntu 16.04 上 CUDA_10.0及cuDNN的安装
  • cudnn-8.0/9.0/10.0-linux-x64-v6.0/7.0/7.1/7.2/7.3/7.4.tgz下载
-------- 本文结束 感谢阅读 --------
相关文章
  • KODExplorer可道云安装过程中遇到的一些问题
  • Ubuntu下安装配置Anaconda和JupyterNoteBook环境
  • Docker:手动下载安装Docker镜像
  • Docker:基于base_centos安装SuperSet
  • Docker:基于base_centos安装MySQL
觉得文章写的不错的话,请我喝瓶怡宝吧!😀
SiriYang 微信支付

微信支付

SiriYang 支付宝

支付宝

  • 本文标题: Ubuntu下安装NVIDIA驱动及CUDA机器学习环境
  • 本文作者: SiriYang
  • 创建时间: 2020年09月28日 - 11时09分
  • 修改时间: 2021年10月29日 - 18时10分
  • 本文链接: https://blog.siriyang.cn/posts/20200928111809id.html
  • 版权声明: 本博客所有文章除特别声明外,均采用 CC BY-NC-SA 4.0 许可协议。转载请注明出处!
Linux 环境 Ubuntu
Ubuntu下安装配置Anaconda和JupyterNoteBook环境
Linux基础操作笔记
SiriYang

SiriYang

努力搬砖攒钱买镜头的摄影迷
320 日志
33 分类
88 标签
RSS
GitHub E-Mail
Creative Commons
Links
  • 友情链接
  • 作品商铺

蜀ICP备19008337号 © 2019 – 2025 SiriYang | 1.7m | 25:41
0%