Learn from the blog: address
1. Install NVIDIA graphics driver
1. Shield nouveau driver
编写（如果没有会新建），下面两个文件
# vi /etc/modprobe.d/nvidia-installer-disable-nouveau.conf
# vi /lib/modprobe.d/nvidia-installer-disable-nouveau.conf
在两个文件下写入如下内容
blacklist nouveau
options nouveau modeset=0
2. Redo the initramfs image
① Make a backup of the current image
cp /boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r).img.bak
②Create a new mirror
dracut /boot/initramfs-$(uname -r).img $(uname -r)
③Restart
reboot
3. Pre-installed components (networking)
yum install gcc kernel-devel kernel-headers
4. Download the official NVIDIA driver according to the current system kernel.
Address: https://www.nvidia.cn/Download/index.aspx?lang=cn#
5. Install the NVIDIA driver and
perform the following installation steps. The kernel source path must be specified, otherwise An error will be reported; the version of the kernel is related to the system kernel, there may be differences
①Locate to the directory where the driver is downloaded
②Execute
./NVIDIA-Linux-x86_64-450.66.run --kernel-source-path=/usr/src/kernels/3.10.0-1127.19.1.el7.x86_64 -k $(uname -r)
Description:
-
NVIDIA-Linux-x86_64-450.66.run is the downloaded driver;
-
3.10.0-1127.19.1.el7.x86_64 is the system kernel source code viewed through the ll /usr/src/kernels/ command
6, installation and decompression
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64 384.81.......................................
..................................................................
..................................................................
Regular selection
Install 32-bit compatibility library -yes
PS: Most of the default options are fine, some are options for post-interaction with NVIDIA, and the default is fine.
7. Verification
execution:
lspci |grep NVIDIA
nvidia-smi
The following effects indicate successful installation
Second, install the CUDA driver
1. Download the CUDA
address: https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=CentOS&target_version=7&target_type=rpmnetwork
According to the provided address, download to the local
2. Installation
Installation execution:
./cuda_11.0.3_450.51.06_linux.run
At this time, there may be insufficient permissions, and authorization is required. Here I directly give the highest permissions
chmod 777 cuda_11.0.3_450.51.06_linux.run
After the agreement appears, choose to accept:
Installation:
There is no need to install the graphics card driver (the screenshot here is wrong) and
a warning message appears, which has no effect on the installation.
Screenshot log:
[root@host83 masr]# sudo sh cuda_11.0.3_450.51.06_linux.run
===========
= Summary =
===========
Driver: Not Selected
Toolkit: Installed in /usr/local/cuda-11.0/
Samples: Installed in /root/, but missing recommended libraries
Please make sure that
- PATH includes /usr/local/cuda-11.0/bin
- LD_LIBRARY_PATH includes /usr/local/cuda-11.0/lib64, or, add /usr/local/cuda-11.0/lib64 to /etc/ld.so.conf and run ldconfig as root
To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-11.0/bin
***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least .00 is required for CUDA 11.0 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 /var/log/cuda-installer.log
You can also enter the log to view (location is as follows):
/var/log/cuda-installer.log
3. Configure environment variables
vi /etc/profile
Enter the file, add configuration
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
Make environment variables take effect immediately
source /etc/profile ;
The installation is complete!
If you help, remember to like it~