ubuntu安装cuda和驱动

cuda安装教程

查看是否有自带驱动(安装出现自带驱动报错时执行)

lsmod | grep nouveau
没有输出就行  

若果有输出
打开系统的blacklist
sudo vim /etc/modprobe.d/blacklist-nouveau.conf

在blacklist中添加禁用nouveau的语句
blacklist nouveau
options nouveau modeset=0

更新initramfs
sudo update-initramfs -u

重启系统
sudo reboot

重新进入系统后，检验nouveau是否被禁用
lsmod | grep nouveau

安装必要依赖

gcc g++ make
apt update
apt install build-essential
(build-essential包含了gcc/g++/make,你也可以apt install gcc  apt install g++   apt install make)

检查 kernels 版本是否匹配
报错信息：ERROR: Unable to find the kernel source tree for the currently running kernel.  Please make sure you have installed the kernel source files for your kernel and that they are properly configured; on Red Hat Linux systems, for example, be sure you have the 'kernel-source' or 'kernel-devel' RPM installed.  If you know the correct kernel source files are installed, you may specify the kernel source path with the '--kernel-source-path' command line option.

先进入相应目录查看是否有相应内核：

cd /usr/src/kernels/ && ls -al

如果为空表示没有，需要自己下载：

不要直接 yum install kernel-devel

因为这样下载下来的内核版本不一定匹配，所以实际的下载命令应该为：

yum install -y kernel-devel-$(uname -r) kernel-headers-$(uname -r) 

下载cuda

https://developer.nvidia.com/cuda-12-0-0-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=runfile_local
自己找对应的版本型号

以Linux Ubuntu 20.04 x86_64为例
wget https://developer.download.nvidia.com/compute/cuda/12.0.0/local_installers/cuda_12.0.0_525.60.13_linux.run
下载完成后
sudo sh cuda_12.0.0_525.60.13_linux.run

之后会跳出
End User License Agreement
-------C
NVIDIA Software License Agreement and CUDA Supplement toSoftware License Agreement .
Preface
------
The Software License Agreement in Chapter 1 and the Supplementin Chapter 2 contain license terms and conditions that governthe use of NVIDIA software. By accepting this agreement, youagree to comply with all the terms and conditions applicableto the product(s) included herein.
NVIDIA Driver
Do you accept the above EULA? (accept/decline/quit):
accept
输入accept

CUDA Installer
-[X] Driver
	[X] 450.51.06
+[X] CUDA Toolkit 11.0
 [X] CUDA Samples 11.0
 [X] CUDA Demo Suite 11.0
 [X] CUDA Documentation 11.0
Options
Install
如果输入nvidia-smi有输出表格结果就将Driver取消勾选，如果没有就直接选择install

完成之后会跳出安装完成信息之后需要配置环境变量信息

刷新环境变量，输入： source ~/.bashrc 

测试

测试cuda是否安装成功
nvcc -V
    
驱动装完都把持久化打开
nvidia-smi -pm 1 #打开持久化

后续问题与解决

conda activate环境时报错坏的解释器，是因为将存放anaconda的包文件给移动了，需要vim进入bin文件下的conda包中，修改头文件的conda地址为目前的地址就可以正常使用了

环境变量
c

修改环境变量 PATH 和 LD_LIBRARY_PATH. 在 ~/.bashrc 文件中写入

vim ~/.bashrc
export PATH=/usr/local/cuda-12.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-12.0/lib64:/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH

source ~/.bashrc

/usr/local/cuda-11.7
根据版本号自行修改

最后校验nvcc -V
如果有回显就表明cuda安装没有问题
nvidia-smi有回显证明驱动没问题