How To Install CUDA 9.2 in Ubuntu

September 16^th 2018

1. Download CUDA 9.2
   1. Use this website
   2. 17.04 version still works
2. Install latest NVIDIA graphics drivers
   1. Do NOT use the Nvidia website’s .run file. It will crash
   2. Run the command

      sudo add-apt-repository ppa:graphics-drivers/ppa

      to add the repository
   3. Go into the software center → Software & Updates → Additional Drivers
   4. Select “Using NVIDIA driver metapackage from nvidia-driver-396” and apply changes
      • 396 is the latest, short-term release driver. You have to use the latest short-term release driver instead of the latest long-term release driver for CUDA
   5. Eventually it will prompt you with a EULA and a secureboot password. Remember this password. It will only be used once, and there is a probably a long process if you accidentally type it in wrong
   6. Reboot your computer
   7. A blue screen will appear before the boot sequence. Hit Enroll → Continue, enter in the previously mentioned password, and reboot.
3. Verify that the drivers work by running

   nvidia-smi

   . It should look like this
   • 
4. Install CUDA “dependencies” with

   sudo apt-get install freeglut3 freeglut3-dev libxi-dev libxmu-dev

5. Install CUDA 9.2
   1. Navigate to where you installed the CUDA downloads
   2. Run the command sudo sh cuda_9.2.148_396.37_linux.run
   3. Read the EULA and other information
   4. Answer the questions this way:
      

      
      You are attempting to install on an unsupported configuration. Do you wish to continue?
      (y)es/(n)o [ default is no ]: y	
      *IMPORTANT*
      Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 396.26?
      (y)es/(n)o/(q)uit: n
      
      Install the CUDA 9.2 Toolkit?
      (y)es/(n)o/(q)uit: y
      			
      Enter Toolkit Location
          [ default is /usr/local/cuda-9.2 ]:
      			
      Do you want to install a symbolic link at /usr/local/cuda?
      (y)es/(n)o/(q)uit: y
      			
      Install the CUDA 9.2 Samples?
      (y)es/(n)o/(q)uit: y
      			
      Enter CUDA Samples Location
          [ default is /home/cindy ]: (optional, keep it in ~)
          

   5. Install the patch with

      sudo sh cuda_9.2.148.1_linux.run

   6. Setup the PATH and LD_LIBRARY_PATH environments with
      

      export PATH=/usr/local/cuda-9.2/bin${PATH:+:${PATH}}
      export LD_LIBRARY_PATH=/usr/local/cuda-9.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

      • NOTE: the above command only works for the current session. To make these commands run after a shutdown, edit ~/.bashrc and add the two lines to the bottom of that file
6. Test CUDA by building the samples
   

   cd ~
   mkdir cuda-testing
   cp a NVIDIA_CUDA-9.2_Samples cuda-testing
   cd cuda-testing
   make -j4

7. Run samples
   

   cd bin/x86_64/linux/release
   ./simplePrintf

   and run any others you’d like!

Credit goes to Dr. Donald Kinghorn with this article