Environment Setup 01-How to View Graphics Card Information and Install NVDIA Graphics Driver in Ubuntu
The original post is a bit old, here is an update, it can be regarded as a quick installation note!
https://blog.csdn.net/tanmx219/article/details/86553485
1. Check the graphics card model
Enter the following command:
lshw -c video
The following information will be displayed
WARNING: you should run this program as super-user.
*-display
description: VGA compatible controller
product: GM107M [GeForce GTX 950M]
vendor: NVIDIA Corporation
physical id: 0
bus info: pci@0000:04:00.0
version: a2
width: 64 bits
clock: 33MHz
capabilities: vga_controller bus_master cap_list rom
configuration: driver=nouveau latency=0
resources: irq:30 memory:fa000000-faffffff memory:d0000000-dfffffff memory:ce000000-cfffffff ioport:ec00(size=128) memory:c0000-dffff
WARNING: output may be incomplete or inaccurate, you should run this program as super-user.
2. Find the driver of your graphics card
Use this command to view the drivers you can use:
ubuntu-drivers devices
The display results are as follows,
matth32@matth32-Thurley:~/下载$ ubuntu-drivers devices
== /sys/devices/pci0000:00/0000:00:07.0/0000:04:00.0 ==
modalias : pci:v000010DEd0000139Asv00001025sd0000091Bbc03sc00i00
vendor : NVIDIA Corporation
model : GM107M [GeForce GTX 950M]
driver : nvidia-driver-440 - distro non-free recommended
driver : nvidia-driver-390 - distro non-free
driver : nvidia-driver-435 - distro non-free
driver : xserver-xorg-video-nouveau - distro free builtin
Go to the NVDIA driver search page to search for the driver model required by the graphics card and download it (as shown in the figure).
Choose the recommended driver,
nvidia-driver-440 - distro non-free recommended
Install after downloading
sudo sh NVIDIA-Linux-x86_64-440.100.run
Appendix 1: How to uninstall Nouveau.
How to remove Nouveau kernel driver (fix Nvidia install error)
This is because, before the installation, Nouveau may have been started, so Nvidia cannot be installed successfully.
#---open a terminal---
sudo apt-get remove nvidia*
sudo apt autoremove
sudo apt-get install dkms build-essential linux-headers-generic
sudo vim /etc/modprobe.d/blacklist.conf
#---save the following info into file blacklist.conf---
blacklist nouveau
blacklist lbm-nouveau
options nouveau modeset=0
alias nouveau off
alias lbm-nouveau off
#---end of the info saved----
#---go back to the terminal---
echo options nouveau modeset=0 | sudo tee -a /etc/modprobe.d/nouveau-kms.conf
sudo update-initramfs -u
reboot
————————————————
参考链接：https://blog.csdn.net/tanmx219/java/article/details/86553485
Appendix 2: Problems encountered during installation
The distribution-provided pre-install script failed! Are you sure you want
to continue?
Continue installation Abort installation
----> Continue installation
Would you like to register the kernel module sources with DKMS? This will
allow DKMS to automatically build a new module, if you install a different
kernel later.
Yes No
---->No
The CC version check failed:
The kernel was built with gcc version 7.4.0 (Ubuntu 7.4.0-1ubuntu1~18.04.1),
but the current compiler version is cc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0.
This may lead to subtle problems; if you are not certain whether the
mismatched compiler will be compatible with your kernel, you may wish to
abort installation, set the CC environment variable to the name of the
compiler used to compile your kernel, and restart installation.
Ignore CC version check Abort installation
----> select ignore
WARNING: Ignoring CC version mismatch:
The kernel was built with gcc version 7.4.0 (Ubuntu
7.4.0-1ubuntu1~18.04.1), but the current compiler version is cc
(Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0.
OK
----> OK
Install NVIDIA's 32-bit compatibility libraries?
Yes No
----> No
An incomplete installation of libglvnd was found. All of the essential
libglvnd libraries are present, but one or more optional components are
missing. Do you want to install a full copy of libglvnd? This will overwrite
any existing libglvnd libraries.
Don't install Install and overw Abort installation.
---->Install and overw
Would you like to run the nvidia-xconfig utility to automatically update
your X configuration file so that the NVIDIA X driver will be used when you
restart X? Any pre-existing X configuration file will be backed up.
Yes No
---->Yes
Your X configuration file has been successfully updated. Installation of
the NVIDIA Accelerated Graphics Driver for Linux-x86_64 (version: 440.100)
is now complete.
OK
----> OK
Three, check after installation
Check the NVIDIA version to check whether the installation is complete
$ nvidia-smi
Environment Build 02-Install cuda and cudnn on Ubuntu
Check the installed nvidia driver version number
nvidia-smi //第一行信息
~$ nvidia-smi
Fri Jul 24 11:55:29 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.100 Driver Version: 440.100 CUDA Version: 10.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 950M Off | 00000000:04:00.0 On | N/A |
| N/A 35C P8 N/A / N/A | 189MiB / 4038MiB | 1% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1129 G /usr/lib/xorg/Xorg 12MiB |
| 0 1179 G /usr/bin/gnome-shell 47MiB |
| 0 3130 G /usr/lib/xorg/Xorg 59MiB |
| 0 3281 G /usr/bin/gnome-shell 62MiB |
| 0 3990 G /usr/lib/firefox/firefox 1MiB |
+-----------------------------------------------------------------------------+
You can see that the version is 10.2.
Select the installed CUDA version and download
Download
link NVIDIA's official cuda and driver correspondence
Find a button similar to DOWNLOAD－"LEGACY RELEASES, and finally find the version you need
Install cuda
Use the installation method provided by the official website
Base Installer
Installation Instructions:
wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
sudo sh cuda_10.2.89_440.33.01_linux.run
If the network is not good, it will be a bit slow, 2G more. Of course, you can also install it after downloading the tool.
During installation, you will be asked if you want to accept anything, enter accept and then press Enter to install.
Under normal circumstances, if the installation is successful, the following prompt words will be given
===========
= Summary =
===========
Driver: Installed
Toolkit: Installed in /usr/local/cuda-10.2/
Samples: Installed in /home/matthew/, but missing recommended libraries
Please make sure that
- PATH includes /usr/local/cuda-10.2/bin
- LD_LIBRARY_PATH includes /usr/local/cuda-10.2/lib64, or, add /usr/local/cuda-10.2/lib64 to /etc/ld.so.conf and run ldconfig as root
To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-10.2/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall
Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-10.2/doc/pdf for detailed information on setting up CUDA.
Logfile is /var/log/cuda-installer.log
If the installation is unsuccessful, it is mostly because drm is enabled in the graphical interface. Use the following method to switch to text console mode, and then use the installation command
sudo sh cuda_10.2.89_440.33.01_linux.run
Of course, if you are using the character interface for the first time, it is best to set a password first, so as not to be unable to log in.
sudo passwd root
Then use the following method to install in the character interface. After installation, return to the graphical interface.
When an error is reported during installation-switch to the character interface
报错 ERROR: An NVIDIA kernel module ‘nvidia-drm’ appears to already be loaded in your kernel.
Close the graphical user interface
sudo systemctl set-default multi-user.target
sudo reboot
Open the graphical user interface
sudo systemctl set-default graphical.target
sudo reboot
Generally speaking, after switching to the character interface, there is no problem with the installation. If it is difficult to switch to the graphical interface, you can consider the following methods
(1)edit /etc/default/grub
GRUB_CMDLINE_LINUX_DEFAULT="text"
GRUB_TERMINAL=console
(2)After saving changes you need to run:
sudo update-grub
sudo systemctl enable multi-user.target --force
sudo systemctl set-default multi-user.target
(3)Undoing text-mode
sudo systemctl enable graphical.target --force
sudo systemctl set-default graphical.target
Configure Cuda environment variables after installation
After the installation is successful in the text console mode, there will be a prompt, we can configure it according to the prompt, here we configure it according to the practice of most people on the Internet,
sudo gedit ~/.bashrc is added
at the end of ~/.bashrc (a total of three additions have been seen online): the
first
export PATH=/usr/local/cuda-10.2/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.2.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
The second
export PATH=/usr/local/cuda-10.2/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda10.2/lib64
The third type (* $PATH means adding to PATH)
export PATH=/usr/local/cuda-10.2/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib64:$LD_LIBRARY_PATH
export CUDA_HOME=/usr/local/cuda
After configuring the environment variable, you must update it, otherwise it will not take effect immediately. You can also restart the computer to make the environment variables take effect:
$source ~/.bashrc
or:
$source /etc/profile
source ~/.bashrc
nvcc --version
If the following text is displayed, the installation is successful:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Wed_Oct_23_19:24:38_PDT_2019
Cuda compilation tools, release 10.2, V10.2.89
CUDNN installation
To download cuDNN,
first go to the official website to download cuDNN (https://developer.nvidia.com/cudnn), you need to be registered as a developer to download, and direct downloading may be very slow. Note the choice of version. For example, I chose
libcudnn7_7.6.5.32-1+cuda10.2_amd64.deb
corresponding to
Download cuDNN v7.6.5 (November 18th, 2019), for CUDA 10.2
Library for Windows, Mac, Linux, Ubuntu and RedHat /Centos(x86_64architecture)
cuDNN Library for Windows 7
cuDNN Library for Windows 10
cuDNN Library for Linux
cuDNN Runtime Library for Ubuntu18.04 (Deb) －－－> Select this package
cuDNN Developer Library for Ubuntu18.04 (Deb)
cuDNN Code Samples and User Guide for Ubuntu18.04 (Deb)
cuDNN Runtime Library for Ubuntu16.04 (Deb)
cuDNN Developer Library for Ubuntu16.04 (Deb)
cuDNN Code Samples and User Guide for Ubuntu16.04 (Deb)
After downloading, right click "open with software installation"-"install", and then OK.
Of course, you can also use the copy method to open the DEB package, there is a data package in it, and the files in it are all placed according to the path. The official installation method is
-
Navigate to your directory containing the cuDNN Tar file.
-
Unzip the cuDNN package.
$ tar -xzvf cudnn-10.2-linux-x64-v7.6.5.32.tgz -
Copy the following files into the CUDA Toolkit directory, and change the file permissions.
$ sudo cp cuda/include/cudnn.h /usr/local/cuda/include $ sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64 $ sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
To repeat, it is mainly to decompress and then execute
sudo cp cuda/include/cudnn.h /usr/local/cuda/include/
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/
Manual copy also needs to set permissions
sudo chmod a+r /usr/local/cuda/include/cudnn.h
sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
Check the cudnn version
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
If the following text is displayed, the installation is successful:
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 5
--
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)
#include "driver_types.h"
Reference materials:
「1」：https://blog.csdn.net/qiancaobaicheng/article/details/95096354