Ubtuntu16.04 CUDA安装

最新推荐文章于 2024-03-18 16:45:22 发布

霜桥月馆

最新推荐文章于 2024-03-18 16:45:22 发布

阅读量2.3k

点赞数

分类专栏： Linux LInux问题与环境配置

本文链接：https://blog.csdn.net/renhaofan/article/details/80969344

版权

Linux 同时被 2 个专栏收录

77 篇文章 2 订阅

订阅专栏

LInux问题与环境配置

77 篇文章 3 订阅

订阅专栏

CUDA安装官方教程
 CUDNN安装官方教程

1. 安装驱动(可选)

有用的驱动调试命令https://blog.csdn.net/wf19930209/article/details/81877822,
可以参考这里，https://blog.csdn.net/ghw15221836342/article/details/79571559。
如果选择这里安装了驱动，那么在第二步安装CUDA的时候就不要选择安装驱动了,
注:因为我是这里安装的，所以最后我的信息显示的是
CUDA Driver Version = 9.1, CUDA Runtime Version = 8.0
如果你想方便安装可以在安装CUDA的时候:
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 375.26? (y)es/(n)o/(q)uit: yes
那么最后你测试程序输出的信息应该是
CUDA Driver Version = 9.1, CUDA Runtime Version = 8.0

2. 安装CUDA tool kit 8.0

cuda下载网址

方法一：下载deb安装包

$ sudo dpkg -i cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb

但是我这样报错了

dpkg-deb: error: 'cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb' is not a debian format archive
dpkg: error processing archive cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb (--install):
 subprocess dpkg-deb --control returned error exit status 2
Errors were encountered while processing:
 cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb

我查了一下试了一下

$ sudo dpkg -i --force-overwrite cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb

但是并没有什么用。blog说是因为deb里似乎带了旧显卡驱动，安装完后还要重新装显卡驱动。可能是因为我安装了的驱动版本过高，而且没有卸载旧的驱动版本导致的。

方法二：下载 runfile安装包

$ sudo sh cuda_8.0.61_375.26_linux.run

安装过程出现如下选择，如果你在安装cuda之前安装了驱动，要选择no，如果没有选择yes
安装过程中会询问是否下载Samples，建议选择，方便安装后做测试

Do you accept the previously read EULA?
accept/decline/quit: accept

Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 375.26?
(y)es/(n)o/(q)uit: no

Install the CUDA 8.0 Toolkit?
(y)es/(n)o/(q)uit: yes

Enter Toolkit Location
 [ default is /usr/local/cuda-8.0 ]: 

Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: 
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: yes

Install the CUDA 8.0 Samples?
(y)es/(n)o/(q)uit: yes

Enter CUDA Samples Location
 [ default is /home/ai ]: 

Installing the CUDA Toolkit in /usr/local/cuda-8.0 ...

Missing recommended library: libXmu.so

Installing the CUDA Samples in /home/ai ...
Copying samples to /home/ai/NVIDIA_CUDA-8.0_Samples now...
Finished copying samples.

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

Driver:   Not Selected
Toolkit:  Installed in /usr/local/cuda-8.0
Samples:  Installed in /home/ai, but missing recommended libraries

Please make sure that
 -   PATH includes /usr/local/cuda-8.0/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-8.0/lib64, or, add /usr/local/cuda-8.0/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run the uninstall script in /usr/local/cuda-8.0/bin

Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-8.0/doc/pdf for detailed information on setting up CUDA.

***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least 361.00 is required for CUDA 8.0 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 /tmp/cuda_install_10888.log

配置环境变量

配置一些环境变量，自我感觉可有可无，但是许多的方案都会有所涉及，包括安装的输出信息，建议配上吧。

$ sudo vim /etc/profile # 文件末尾添加一下两行
$ export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}}
$ export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
$ sudo gedit /etc/profile #导入动态库 
$ export PATH = /usr/local/cuda/bin:$PATH #在文件末尾加入
$ sudo gedit /etc/ld.so.conf.d/cuda.conf #创建链接文件 
$ /usr/local/cuda/lib64 #在打开的文件中添加以下内容
$ sudo ldconfig #执行使得链接生效

检查安装是否成功

$ cd NVIDIA_CUDA-8.0_Samples/1_Utilities/deviceQuery
$ make
$ sudo ./deviceQuery

注意关注最后一行，pass就没错
输出如下:

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.1, CUDA Runtime Version = 8.0, NumDevs = 1, Device0 = GeForce GTX 960M
Result = PASS

安装cudnnv5.1

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

$ tar -zxvf xxx.tar
$ cd xxx/include   #切换进入include路径 
$ sudo cp cudnn.h /usr/local/cuda/include/    #复制头文件
$ cd xxx/lib64    #切换进入lib64目录
$ sudo cp lib* /usr/local/cuda/lib64/          #复制动态链接库
$ cd /usr/local/cuda/lib64/
$ sudo rm -rf libcudnn.so libcudnn.so.5        #删除原有动态文件
$ sudo ln -sf libcudnn.so.5.0.5 libcudnn.so.5  #生成软衔接
$ sudo ln -sf libcudnn.so.5 libcudnn.so      #生成软链接
$ sudo ldconfig

在本地cuda/lib64下检测一下

$ ll

有类似这样的：

lrwxrwxrwx  1 root root        13 Oct 15 06:55 libcudnn.so -> libcudnn.so.7*
lrwxrwxrwx  1 root root        17 Oct 15 06:55 libcudnn.so.7 -> libcudnn.so.7.3.1*
-rwxr-xr-x  1 root root 302770160 Oct 15 06:53 libcudnn.so.7.3.1*

说明安装成功了