View GPU Model:
lspci | grep -i nvidia
Driver Installation:
https://www.nvidia.cn/Download/index.aspx?lang=cn
Download the corresponding version of the driver for driving;
Install gcc, g ++, make the latest version of the program;
Installation ubuntu 32 bit compatibility libraries;
Run .run file to install it; read the information and tips for installation, pay attention and not to install opengl, online mentions pit, I chose not to install the file opengl related, of course, you can also try;
./nvidiainstaller.run --no-opengl-files
After the restart, check whether the nvidia-settings to run, run and check the information;
cuda installation:
Download the latest file cuda: https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=runfilelocal
Note that the corresponding version to install;
./cuda-installer.run
After installation, remember to add environment variables:
we ~ / .bashrc
Add at the end of file
export PATH="/usr/local/cuda-10.1/bin:$PATH"
export LD_LIBRARY_PATH="/usr/lcoal/cuda-10.1/lib64:$LD_LIBRARY_PATH"
Finally, to take effect
source ~/.bashrc
After installing the driver and cuda, execute test routines, the measuring environment is installed successfully:
/usr/local/cuda-10.1/samples/1_Utilities/deviceQuery cuda sample, make use make to compile and run. After the execution, cuda display by representing a successful installation;
Detailed installation tutorial, please refer to the User article: https://www.cnblogs.com/zmbreathing/p/CUDA.html , I have been verified, the installation process is no problem;
Stay updated, if helpful to you, please pay attention to cnblogs.com/xuyaowen, for more information;