随笔- 65 文章- 0 评论- 1 阅读- 10185

Cuda 驱动安装

Author: ACatSmiling

Since: 2024-11-13

CUDA（Compute Unified Device Architecture）：是 NVIDIA 推出的一种并行计算平台和编程模型，它允许开发者利用 NVIDIA GPU（图形处理器）的强大计算能力进行通用计算，而不仅仅局限于图形处理。简单来说，CUDA 提供了一种方式，让程序员可以像编写 CPU 程序一样来编写能够在 GPU 上高效运行的程序。

安装步骤：

根据操作系统版本和 GPU 型号等，下载对应的 cuda 驱动，下载地址：https://developer.nvidia.com/cuda-downloads
检查服务器 GPU 识别情况，确保 GPU 卡能够完全识别，如不能识别，需要进行重新插拔、对调测试等步骤进行硬件排查。
```
 [root@desong-gpu-wushan ~]# lspci | grep -i nvidia
00:07.0 3D controller: NVIDIA Corporation Device 2236 (rev a1)
```

安装 gcc，g++ 编译器。（GPU 驱动安装时，需要 gcc 编译器）

 [root@desong-gpu-wushan ~]# gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.8.5/lto-wrapper
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-linker-hash-style=gnu --enable-languages=c,c++,objc,obj-c++,java,fortran,ada,go,lto --enable-plugin --enable-initfini-array --disable-libgcj --with-isl=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/isl-install --with-cloog=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/cloog-install --enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
Thread model: posix
gcc version 4.8.5 20150623 (Red Hat 4.8.5-16) (GCC) 
[root@desong-gpu-wushan ~]# g++ -v
Using built-in specs.
COLLECT_GCC=g++
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.8.5/lto-wrapper
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-linker-hash-style=gnu --enable-languages=c,c++,objc,obj-c++,java,fortran,ada,go,lto --enable-plugin --enable-initfini-array --disable-libgcj --with-isl=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/isl-install --with-cloog=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/cloog-install --enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
Thread model: posix
gcc version 4.8.5 20150623 (Red Hat 4.8.5-16) (GCC)

yum install gcc和yum install gcc-c++。

安装 kernel-devel 和 kernel-headers 软件包。

 [root@desong-gpu-wushan ~]# yum list | grep kernel
Repodata is over 2 weeks old. Install yum-cron? Or run: yum makecache fast
kernel.x86_64                             3.10.0-693.el7               @anaconda
kernel-abi-whitelists.noarch              3.10.0-693.el7               @base-52 
kernel-devel.x86_64                       3.10.0-693.el7               @anaconda
kernel-headers.x86_64                     3.10.0-693.el7               @anaconda
kernel-tools.x86_64                       3.10.0-693.el7               @anaconda
kernel-tools-libs.x86_64                  3.10.0-693.el7               @anaconda
abrt-addon-kerneloops.x86_64              2.1.11-60.el7.centos         base     
erlang-kernel.x86_64                      R16B-03.18.el7               epel     
kernel.x86_64                             3.10.0-1160.42.2.el7         updates  
kernel-abi-whitelists.noarch              3.10.0-1160.42.2.el7         updates  
kernel-debug.x86_64                       3.10.0-1160.42.2.el7         updates  
kernel-debug-devel.x86_64                 3.10.0-1160.42.2.el7         updates  
kernel-devel.x86_64                       3.10.0-1160.42.2.el7         updates  
kernel-doc.noarch                         3.10.0-1160.42.2.el7         updates  
kernel-headers.x86_64                     3.10.0-1160.42.2.el7         updates  
kernel-tools.x86_64                       3.10.0-1160.42.2.el7         updates  
kernel-tools-libs.x86_64                  3.10.0-1160.42.2.el7         updates  
kernel-tools-libs-devel.x86_64            3.10.0-1160.42.2.el7         updates  
libreport-plugin-kerneloops.x86_64        2.1.11-53.el7.centos         base     
lirc-disable-kernel-rc.noarch             0.10.0-16.el7                epel     
php-symfony-http-kernel.noarch            2.8.12-2.el7                 epel     
texlive-l3kernel.noarch                   2:svn29409.SVN_4469-45.el7   base     
texlive-l3kernel-doc.noarch               2:svn29409.SVN_4469-45.el7   base

yum install kernet-devel-$(uname -r) kernel-headers-$(uname -r)。

禁用系统自带的 nouveau 模块。

 # 检查 nouveau 模块是否加载
[root@desong-gpu-wushan ~]# lsmod | grep nouveau
nouveau              1622010  0 
mxm_wmi                13021  1 nouveau
wmi                    19070  2 mxm_wmi,nouveau
video                  24520  1 nouveau
i2c_algo_bit           13413  1 nouveau
drm_kms_helper        159169  2 cirrus,nouveau
ttm                    99345  2 cirrus,nouveau
drm                   370825  5 ttm,drm_kms_helper,cirrus,nouveau
i2c_core               40756  5 drm,i2c_piix4,drm_kms_helper,i2c_algo_bit,nouveau
# 编辑 /usr/lib/modprobe.d/dist-blacklist.conf，添加以下内容，如果没有则创建
[root@desong-gpu-wushan modprobe.d]# vim /usr/lib/modprobe.d/dist-blacklist.conf
blacklist nouveau
options nouveau modeset=0
# 使内核生效
[root@desong-gpu-wushan modprobe.d]# dracut -force

修改系统运行级别为文本模式，GPU 驱动安装必须在文本模式下进行。

 [root@desong-gpu-wushan modprobe.d]# systemctl set-default multi-user.target
Removed symlink /etc/systemd/system/default.target.
Created symlink from /etc/systemd/system/default.target to /usr/lib/systemd/system/multi-user.target.
# 重启系统
[root@desong-gpu-wushan modprobe.d]# reboot

验证 nouveau 模块禁用生效，并进入文本模式。
```
 [root@desong-gpu-wushan ~]# lsmod | grep nouveau
```

cuda 安装。

 # 赋予 cuda 安装包可执行权限
[root@desong-gpu-wushan ~]# chmod +x cuda_12.2.0_535.54.03_linux.run
# 执行 cuda 安装命令
[root@desong-gpu-wushan ~]# ./cuda_12.2.0_535.54.03_linux.run --no-opengl-libs

默认全部安装，如果已经安装过 GPU 驱动，则取消 Driver 选项的勾选。

 Driver:   Installed
Toolkit:  Installed in /usr/local/cuda-12.2/
 
Please make sure that
 
 -   PATH includes /usr/local/cuda-12.2/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-12.2/lib64, or, add /usr/local/cuda-12.2/lib64 to /etc/ld.so.conf and run ldconfig as root
 
To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-12.2/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall
Logfile is /var/log/cuda-installer.log

配置环境变量。

 [root@desong-gpu-wushan ~]# vim /etc/profile
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

生效环境变量。

 [root@desong-gpu-wushan ~]# source /etc/profile

测试 cuda 是否安装成功。

 [root@desong-gpu-wushan ~]# nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2023 NVIDIA Corporation
Built on Tue_Jun_13_19:16:58_PDT_2023
Cuda compilation tools, release 12.2, V12.2.91
Build cuda_12.2.r12.2/compiler.32965470_0

修改系统运行级别为图形模式，然后重启系统。

 [root@desong-gpu-wushan ~]# systemctl set-default graphical.target
Removed symlink /etc/systemd/system/default.target.
Created symlink from /etc/systemd/system/default.target to /usr/lib/systemd/system/graphical.target.
[root@desong-gpu-wushan ~]# reboot

验证GPU驱动及CUDA安装成功，以及图形化界面显示正常。

 [root@desong-gpu-wushan ~]# nvidia-smi
Wed Nov  6 07:18:06 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.54.03              Driver Version: 535.54.03    CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA A10                     Off | 00000000:00:07.0 Off |                    0 |
|  0%   39C    P0              58W / 150W |      4MiB / 23028MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|  No running processes found                                                           |
+---------------------------------------------------------------------------------------+

posted @ 2024-11-14 23:14 ACatSmiling 阅读(345) 评论(0) 编辑收藏举报

刷新页面返回顶部

登录后才能查看或发表评论，立即登录或者逛逛博客园首页

相关博文：

· Hyper-V 安装 CentOS

· Linux 的系统信息命令

· 【PyTorch】n卡驱动、CUDA Toolkit、cuDNN全解安装教程

· Ubuntu NVIDIA显卡驱动+CUDA安装(多版本共存)

· 运维系列&AI系列&Lenovo-G双系统系列（前传-装之前搜索的）：ubuntu22.04安装nvidia驱动+cuda11.7+cudnn8.6

阅读排行：
· 全程不用写代码，我用AI程序员写了一个飞机大战
· DeepSeek 开源周回顾「GitHub 热点速览」
· MongoDB 8.0这个新功能碉堡了，比商业数据库还牛
· 记一次.NET内存居高不下排查解决与启示
· 白话解读 Dapr 1.15：你的「微服务管家」又秀新绝活了

公告

邮箱：wdshfut@163.com
Github：https://github.com/ACatSmiling
Gitee：https://gitee.com/ACatSmiling

昵称： ACatSmiling
园龄： 7个月
粉丝： 4
关注： 0

+加关注

2025年3月

日

一

二

三

四

五

六

随笔分类

随笔档案

阅读排行榜

评论排行榜

1. Java 内存模型(1)

未来像盛夏的大雨，在我们还来不及撑开伞时就扑面而来。

Cuda 驱动安装

公告

搜索

常用链接

我的标签

随笔分类

随笔档案

阅读排行榜

评论排行榜

推荐排行榜

最新评论

	[root@desong-gpu-wushan ~]# lspci \| grep -i nvidia
	00:07.0 3D controller: NVIDIA Corporation Device 2236 (rev a1)

	[root@desong-gpu-wushan ~]# gcc -v
	Using built-in specs.
	COLLECT_GCC=gcc
	COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.8.5/lto-wrapper
	Target: x86_64-redhat-linux
	Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-linker-hash-style=gnu --enable-languages=c,c++,objc,obj-c++,java,fortran,ada,go,lto --enable-plugin --enable-initfini-array --disable-libgcj --with-isl=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/isl-install --with-cloog=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/cloog-install --enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
	Thread model: posix
	gcc version 4.8.5 20150623 (Red Hat 4.8.5-16) (GCC)
	[root@desong-gpu-wushan ~]# g++ -v
	Using built-in specs.
	COLLECT_GCC=g++
	COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.8.5/lto-wrapper
	Target: x86_64-redhat-linux
	Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-linker-hash-style=gnu --enable-languages=c,c++,objc,obj-c++,java,fortran,ada,go,lto --enable-plugin --enable-initfini-array --disable-libgcj --with-isl=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/isl-install --with-cloog=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/cloog-install --enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
	Thread model: posix
	gcc version 4.8.5 20150623 (Red Hat 4.8.5-16) (GCC)

	[root@desong-gpu-wushan ~]# yum list \| grep kernel
	Repodata is over 2 weeks old. Install yum-cron? Or run: yum makecache fast
	kernel.x86_64 3.10.0-693.el7 @anaconda
	kernel-abi-whitelists.noarch 3.10.0-693.el7 @base-52
	kernel-devel.x86_64 3.10.0-693.el7 @anaconda
	kernel-headers.x86_64 3.10.0-693.el7 @anaconda
	kernel-tools.x86_64 3.10.0-693.el7 @anaconda
	kernel-tools-libs.x86_64 3.10.0-693.el7 @anaconda
	abrt-addon-kerneloops.x86_64 2.1.11-60.el7.centos base
	erlang-kernel.x86_64 R16B-03.18.el7 epel
	kernel.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-abi-whitelists.noarch 3.10.0-1160.42.2.el7 updates
	kernel-debug.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-debug-devel.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-devel.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-doc.noarch 3.10.0-1160.42.2.el7 updates
	kernel-headers.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-tools.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-tools-libs.x86_64 3.10.0-1160.42.2.el7 updates
	kernel-tools-libs-devel.x86_64 3.10.0-1160.42.2.el7 updates
	libreport-plugin-kerneloops.x86_64 2.1.11-53.el7.centos base
	lirc-disable-kernel-rc.noarch 0.10.0-16.el7 epel
	php-symfony-http-kernel.noarch 2.8.12-2.el7 epel
	texlive-l3kernel.noarch 2:svn29409.SVN_4469-45.el7 base
	texlive-l3kernel-doc.noarch 2:svn29409.SVN_4469-45.el7 base

	# 检查 nouveau 模块是否加载
	[root@desong-gpu-wushan ~]# lsmod \| grep nouveau
	nouveau 1622010 0
	mxm_wmi 13021 1 nouveau
	wmi 19070 2 mxm_wmi,nouveau
	video 24520 1 nouveau
	i2c_algo_bit 13413 1 nouveau
	drm_kms_helper 159169 2 cirrus,nouveau
	ttm 99345 2 cirrus,nouveau
	drm 370825 5 ttm,drm_kms_helper,cirrus,nouveau
	i2c_core 40756 5 drm,i2c_piix4,drm_kms_helper,i2c_algo_bit,nouveau

	# 编辑 /usr/lib/modprobe.d/dist-blacklist.conf，添加以下内容，如果没有则创建
	[root@desong-gpu-wushan modprobe.d]# vim /usr/lib/modprobe.d/dist-blacklist.conf
	blacklist nouveau
	options nouveau modeset=0

	# 使内核生效
	[root@desong-gpu-wushan modprobe.d]# dracut -force

	[root@desong-gpu-wushan modprobe.d]# systemctl set-default multi-user.target
	Removed symlink /etc/systemd/system/default.target.
	Created symlink from /etc/systemd/system/default.target to /usr/lib/systemd/system/multi-user.target.
	# 重启系统
	[root@desong-gpu-wushan modprobe.d]# reboot

	# 赋予 cuda 安装包可执行权限
	[root@desong-gpu-wushan ~]# chmod +x cuda_12.2.0_535.54.03_linux.run
	# 执行 cuda 安装命令
	[root@desong-gpu-wushan ~]# ./cuda_12.2.0_535.54.03_linux.run --no-opengl-libs

	Driver: Installed
	Toolkit: Installed in /usr/local/cuda-12.2/

	Please make sure that

	- PATH includes /usr/local/cuda-12.2/bin
	- LD_LIBRARY_PATH includes /usr/local/cuda-12.2/lib64, or, add /usr/local/cuda-12.2/lib64 to /etc/ld.so.conf and run ldconfig as root

	To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-12.2/bin
	To uninstall the NVIDIA Driver, run nvidia-uninstall
	Logfile is /var/log/cuda-installer.log

	[root@desong-gpu-wushan ~]# vim /etc/profile
	export PATH=/usr/local/cuda/bin:$PATH
	export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

	[root@desong-gpu-wushan ~]# nvcc -V
	nvcc: NVIDIA (R) Cuda compiler driver
	Copyright (c) 2005-2023 NVIDIA Corporation
	Built on Tue_Jun_13_19:16:58_PDT_2023
	Cuda compilation tools, release 12.2, V12.2.91
	Build cuda_12.2.r12.2/compiler.32965470_0

	[root@desong-gpu-wushan ~]# systemctl set-default graphical.target
	Removed symlink /etc/systemd/system/default.target.
	Created symlink from /etc/systemd/system/default.target to /usr/lib/systemd/system/graphical.target.
	[root@desong-gpu-wushan ~]# reboot

	[root@desong-gpu-wushan ~]# nvidia-smi
	Wed Nov 6 07:18:06 2024
	+---------------------------------------------------------------------------------------+
	\| NVIDIA-SMI 535.54.03 Driver Version: 535.54.03 CUDA Version: 12.2 \|
	\|-----------------------------------------+----------------------+----------------------+
	\| GPU Name Persistence-M \| Bus-Id Disp.A \| Volatile Uncorr. ECC \|
	\| Fan Temp Perf Pwr:Usage/Cap \| Memory-Usage \| GPU-Util Compute M. \|
	\| \| \| MIG M. \|
	\|=========================================+======================+======================\|
	\| 0 NVIDIA A10 Off \| 00000000:00:07.0 Off \| 0 \|
	\| 0% 39C P0 58W / 150W \| 4MiB / 23028MiB \| 0% Default \|
	\| \| \| N/A \|
	+-----------------------------------------+----------------------+----------------------+

	+---------------------------------------------------------------------------------------+
	\| Processes: \|
	\| GPU GI CI PID Type Process name GPU Memory \|
	\| ID ID Usage \|
	\|=======================================================================================\|
	\| No running processes found \|
	+---------------------------------------------------------------------------------------+