How to install NVIDIA CUDA on Debian 10
CUDA® is a parallel computing platform and programming model developed by NVIDIA for general computing on graphical processing units (GPUs). With CUDA, developers are able to dramatically speed up computing applications by harnessing the power of GPUs.
We will install NVIDIA CUDA on a Debian 10 server. Run the following commands as user root. If you are logged in with sudo user, you can run the command “sudo su” to become user root.
Add NVIDIA repository
|
1 2 3 |
apt install software-properties-common -y add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/debian10/x86_64/ /" apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/debian10/x86_64/7fa2af80.pub |
Enable the contrib repository
|
1 |
add-apt-repository contrib |
Update the apt cache
|
1 |
apt-get --allow-releaseinfo-change update |
Install cuda
|
1 |
apt-get -y install cuda |
To find Nvidia driver version, you can run the command cat /proc/driver/nvidia/version
|
1 2 3 4 |
root@sok-1:~# cat /proc/driver/nvidia/version NVRM version: NVIDIA UNIX x86_64 Kernel Module 470.103.01 Thu Jan 6 12:10:04 UTC 2022 GCC version: gcc version 8.3.0 (Debian 8.3.0-6) root@sok-1:~# |
cuda installed in folder /usr/local/cuda-11.4/bin/, to find version, run
|
1 2 3 4 5 6 7 |
root@sok-1:~# /usr/local/cuda-11.4/bin/nvcc -V nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2021 NVIDIA Corporation Built on Mon_Oct_11_21:27:02_PDT_2021 Cuda compilation tools, release 11.4, V11.4.152 Build cuda_11.4.r11.4/compiler.30521435_0 root@sok-1:~# |
Add the folder to PATH, edit the file
|
1 |
vi ~/.bashrc |
Add following
|
1 2 3 4 |
if [ -d "/usr/local/cuda-11.4/bin/" ]; then export PATH=/usr/local/cuda-11.4/bin${PATH:+:${PATH}} export LD_LIBRARY_PATH=/usr/local/cuda-11.4/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} fi |
Save and exit the file. Reload .bashrc with
|
1 |
source ~/.bashrc |
You can find more info on NVIDIA driver with the command
|
1 |
nvidia-smi |
