Install Nvidia Driver CUDA and cuDNN - Ubuntu.md

Install Nvidia Driver CUDA and cuDNN on Ubuntu

Jan 2nd, 2022

Index

• System Specification Check
• NVIDIA Driver Installation
• CUDA Toolkit Installation
• cuDNN Installation
• Nvidia Documentation

System Specification Check

• Check your system architecture to select correct installers for your platform

  $ uname -m
  $ dpkg --print-architecture

NVIDIA Driver Installation

1. Remove old installation

   $ sudo apt-get purge nvidia-*
   $ sudo apt-get update 
   $ sudo apt-get autoremove # DO NOT skip this line

2. Search for latest version of Nvidia driver

   $ apt search nvidia-driver

3. Install Nvidia libraries

   $ sudo apt install libnvidia-common-<version>
   $ sudo apt install libnividia-gl-<version>

4. Install Nvidia driver

   $ sudo apt install nvidia-driver-<version>

5. Reboot and check for the installation

   $ nvidia-smi

CUDA Toolkit Installation

1. Intsall kernel headers and developement packages for your currently running kernel

   $ sudo apt-get install linux-headers-$(uname -r)

2. Download and install CUDA Toolkit

• CUDA Toolkit from Nvidia Developer
  • Select target platform
  • Recommendation: pick deb [network] option of Installer Type
  • Follow the installation instruction on the download page to install CUDA Toolkit
• To include GDS package with CUDA Toolkit
  • GDS Overview: CUDA GPUDirect Storage

  $ sudo apt-get install nvidia-gds 

3. Setup environment

• Config $PATH variable with following script:

  CUDA_HOME=/usr/local/cuda
  PATH=${CUDA_HOME}/bin${PATH:+:${PATH}}
  LD_LIBRARY_PATH=${CUDA_HOME}/lib64 ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
  export LD_LIBRARY_PATH
  export CUDA_HOME
  export PATH

• Add the script to either:
  • ~/.bashrc for user session usage
  • /etc/profile for system wide usage

4. Setup POWER9

• Check NVIDIA Persistence Daemon

  $ systemctl status nvidia-persistenced

• If it is not loaded

  $ sudo systemctl enable nvidia-persistenced

5. Reboot and check for installation

   $ nvcc --version

cuDNN Installation

1. Download cuDNN:

• Nvidia cuDNN from Nvidia Developer (local installer)
  • NVIDIA Developer Program Membership is required to download
  • Select CUDA matching version and target platform

2. Install cuDNN

• Import CUDA GPG key

  $ sudo dpkg -i <downloaded-file>
  $ sudo apt-key add /var/cudnn-local-repo-*/7fa2af80.pub
  $ sudo apt-get update

• To auto-match version of cuDNN v8 with version of CUDA when installing:

  $ sudo apt-get install libcudnn8
  $ sudo apt-get install libcudnn8-dev 
  $ sudo apt-get install libcudnn8-samples 

Nvidia Documentation

• CUDA Tookit Installation Guide
• cuDNN Installation Guide

My Installation

• Operating System: Ubuntu 20.04 x84_64 (64-bit)
• Architecture: amd64
• GPU: Nvidia GeForce GTX 1050
• Installation with success on: Jan 2nd, 2022

Hi @UTKRISHTPATESARIA ,

You are correct! We can have GSD on RTX 3090 in compatibility mode, but GSD cannot do anything with DMA on RTX 3090. I did not aware of this because I did not use much my RTX 3090. What we will have with sudo apt-get install nvidia-gds are only the GDS packages. Sorry for my previous incorrect information.

https://forums.developer.nvidia.com/t/gpudirect-available-on-ubuntu-18-04/192420/5

Thanks for sharing.

Let me see if I can setup GPU Direct RDMA using some custom build or stuff. That's the last hope for me.