Installation of NVIDIA Driver

Disable Nouveau (default graphic driver on Ubuntu)

1. Open /etc/modprobe.d/blacklist-nouveau.conf and add the following lines:

   blacklist nouveau
   options nouveau modeset=0
   

   Save it (sudo privilege may be required).

2. Execute $sudo update-initramfs -u and reboot system.

Install NVIDIA GPU Driver

(You can download the latest NVIDIA GPU driver (.run file) from http://www.nvidia.com/Download/index.aspx)

1. If graphical login-screen appears, press [Alt] + [Ctrl] + [F1] and login by virtual console (CUI environment).

2. Execute $sudo service lightdm stop to kill X server temporarily.

3. Execute $sudo sh NVIDIA-***.run or $sudo sh NVIDIA-***.run --no-opengl-files (on laptops that have both integrated graphic card and NVIDIA-GPU).

   NOTE: DO NOT run the NVIDIA configuration for X windowing system at the end of the installation of the GPU driver on laptop, since your integrated graphic card will be used to display the desktop. The NVIDIA card will run whenever needed automatically.

4. Reboot system

Install CUDA

(You can download the latest CUDA installer (.run file) from https://developer.nvidia.com/cuda-downloads)

1. Execute $sudo sh cuda_***.run You don't have to install GPU driver contained in CUDA installer as you have already installed the latest one in the previous section.

2. Once the installation completes, export PATH and LD_LIBRARY_PATH according to the installer message. EX. open ~/.bashrc with your favorite text editor and add these two lines:

   export PATH=/usr/local/cuda/bin:$PATH
   export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
   

Some additional packages maybe required in order to fully compile CUDA samples: $sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libgl1-mesa-dev

If you receive a compiler error such as /usr/bin/ld: cannot find -lGL, this command may resolve it:

$sudo ln -s /usr/lib/libGL.so.1 /usr/lib/libGL.so

If CUDA is not active, you can enable it manually running the following lines as root:

1. Create /etc/init.d/cuda_setup_script with the following content:

   #!/bin/bash
   
   /sbin/modprobe nvidia
   
   if [ "$?" -eq 0 ]; then
   # Count the number of NVIDIA controllers found.
   NVDEVS=`lspci | grep -i NVIDIA`
   N3D=`echo "$NVDEVS" | grep "3D controller" | wc -l`
   NVGA=`echo "$NVDEVS" | grep "VGA compatible controller" | wc -l`
   
   N=`expr $N3D + $NVGA - 1`
   for i in `seq 0 $N`; do
   mknod -m 666 /dev/nvidia$i c 195 $i
   done
   
   mknod -m 666 /dev/nvidiactl c 195 255
   
   else
   exit 1
   fi
   
   /sbin/modprobe nvidia-uvm
   
   if [ "$?" -eq 0 ]; then
   # Find out the major device number used by the nvidia-uvm driver
   D=`grep nvidia-uvm /proc/devices | awk '{print $1}'`
   
   mknod -m 666 /dev/nvidia-uvm c $D 0
   else
   exit 1
   fi
   

   (Reference: http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-linux/index.html#runfile-verifications)

2. Execute: $sudo chmod +x /etc/init.d/cuda_setup_script

3. Finally run: $sudo update-rc.d cuda_setup_script defaults

   NOTE: DO NOT specify full path of cuda_setup_script as 2nd argument.