❗ NOTE: for Ubuntu 20.04 NVIDIA drivers
are installed by default. Just check it by nvidia-smi command.
- Install graphics driver for Palit GeForce GTX 1080 Ti GameRock 11GB GDDR5X [NEB108TT15LC-1020G].
- Install CUDA library for all users on Ubuntu 18.04.
My answer on AskUbuntu.com: https://askubuntu.com/a/1097211/672237
- How to Install Nvidia CUDA Toolkit on Ubuntu 18.04 LTS
- How to install CUDA 9.2 on Ubuntu 18.04
- NVIDIA CUDA Installation Guide for Linux
- How to install Nvidia driver in ubuntu 18.04
- Commands for NVIDIA install on Ubuntu 16.04
Go to NVIDIA web-site and obtain the latest driver for your GPU. In my case it is:
Product Type: GeForce
Product Series: GeForce 10 Series
Product: GeForce GTX 1080 Ti
Operating System: Linux 64-bit
Language: English (US)
Press <SEARCH> button and check that the founded driver is
supporting your GPU model in "SUPPORTED PRODUCTS" tab.
Download it. In my case the filename is: NVIDIA-Linux-x86_64-410.78.run
# Change permission to run and execute it
chmod +x NVIDIA-Linux-x86_64-410.78.run
# Before installation install gcc and make packages:
sudo apt install gcc
sudo apt install makeIt is better to run driver installation in the text mode.
For the text mode press 3 keys <Ctrl>+<Alt>+<F3> and login to console.
Most likely you'll have problems with previously installed graphics driver called Nouveau.
# Stop graphical interface
sudo service lightdm stop
# If /tmp directory is too small
df -h /tmp
# Increase /tmp size to 10GB.
sudo umount /tmp
sudo mount -t tmpfs -o size=10485760000,mode=1777 overflow /tmp
# Remove Nouveau driver
sudo apt purge remove xserver-xorg-video-nouveau
# Remove previously installed NVIDIA driver
sudo apt purge nvidia*
sudo apt autoremove
sudo apt autoclean
sudo rm -rf /usr/local/cuda*
# Execute file and answer the questions during installation
sudo ./NVIDIA-Linux-x86_64-410.93.run
# Continue installation
# Install NVIDIA's 32-bit compatibility libraries? — Yes
# Whould you like to run the nvidia-xconfig utility... — Yes
# OK
# Start graphical interface
sudo service lightdm start
# Reboot Ubuntu
sudo reboot
# After reboot check if installation is successful
nvidia-smiYou should see terminal output of Nvidia Drivers:
More checks
# Check again
lsmod | grep nouveau # should be zero output
lsmod | grep nvidia # should be non-zero output
# Another check. {tab} means you should press <Tab> button on your keyboard.
cat /proc/driver/nvidia/gpus/{tab}/information
cat /proc/driver/nvidia/gpus/0000\:17\:00.0/information
Model: GeForce GTX 1080 Ti
IRQ: 63
GPU UUID: GPU-e953c3c5-dba7-029e-ce44-7b51417131df
Video BIOS: 86.02.40.00.48
Bus Type: PCIe
DMA Size: 47 bits
DMA Mask: 0x7fffffffffff
Bus Location: 0000:17:00.0
Device Minor: 0
Blacklisted: No
cat /proc/driver/nvidia/gpus/0000\:65\:00.0/information
Model: GeForce GTX TITAN X
IRQ: 64
GPU UUID: GPU-4ce175fe-a313-4c4a-d620-560e96e8ef2e
Video BIOS: 84.00.45.00.03
Bus Type: PCIe
DMA Size: 40 bits
DMA Mask: 0xffffffffff
Bus Location: 0000:65:00.0
Device Minor: 1
Blacklisted: NoYou should see correct model of your GPU:
Error when executing command nvidia-smi:
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.
Warning: Try to not mix runfile and package manager install methods.
Warning: Install CUDA Toolkit into your Virtual Environment. Otherwise, it could be a conflicts between System and Virtual Env. libraries.
Package manager method (link):
# Install ubuntu-drivers
sudo apt install ubuntu-drivers-common
sudo ubuntu-drivers autoinstall
# Check what driver is valid
sudo ubuntu-drivers devicesAfter ubuntu-drivers installation open menu
Run → Preferences → Software & Updates → tab Ubuntu software→
check Proprietary drivers for devices (restricted) option.
Then update from gui or sudo apt update.
# Pick the right driver and run it
sudo apt install nvidia-driver-430
# Sometimes `--fix-missing` flag is needed
# sudo apt install nvidia-driver-535 --fix-missing
# Install CUDA toolkit not on the System, but into your Virtual Environment
# sudo apt install nvidia-cuda-toolkit # don't install it :-)Runfile method:
# Go to directory with nVidia driver
cd ~/Documents/Install/nVidia Geforce GTX 1080 Ti
# Download graphical driver from nVidia web site and make it executable.
chmod +x NVIDIA-Linux-x86_64-418.43.run
# Install necessary packages
sudo apt install gcc make
# Press <Ctrl>+<Alt>+<F3> and login to console.
# Stop graphical interface
sudo service lightdm stop
sudo apt autoremove
sudo apt autoclean
# Execute file and answer the questions during installation
sudo ./NVIDIA-Linux-x86_64-410.93.run
# Start graphical interface
sudo service lightdm start
# Reboot Ubuntu
sudo reboot
# After reboot check if installation is successful
nvidia-smiReinstall nvidia-docker if necessary.
Troubleshooting:
# Remove Nouveau drivers
apt search nouveau # check if all nouveau drivers are uninstalled
sudo apt purge xserver-xorg-video-nouveau
# You can remove previously installed NVIDIA driver
# sudo apt purge nvidia* # there are nvidia-docker packages thereWarning: Install CUDA Toolkit into your Virtual Environment. Otherwise, it could be a conflicts between System and Virtual Env. libraries.
# Update and upgrade your system
sudo apt update
sudo apt upgrade
# Verify you have a CUDA-capable GPU
lspci | grep -i nvidia
# Verify you have a supported version of Linux
# x86_64 line indicates you are running on a 64-bit system
# which is supported by cuda 10.0
uname -m && cat /etc/*release
# View linux kernel header
uname -r # somethins like — 4.15.0-39-generic
# Install gcc, kernel headers and development libraries
sudo apt install gcc-6 # gcc compiler
sudo apt install g++-6 # g++ compiler
sudo apt install linux-headers-$(uname -r) # kernel headers
sudo apt install freeglut3-dev libxmu-dev libpcap-dev # development libsTo download CUDA toolkit open NVIDIA web-site
Select: Linux → x86_64 → Ubuntu → 18.04 → runfile (local).
Download 2.0 GB file: cuda_10.0.130_410.48_linux.run
sudo service lightdm stop # Stop graphical interface
# Change permissions and run it
sudo chmod +x cuda_10.0.130_410.48_linux.run
sudo ./cuda_10.0.130_410.48_linux.run
Do you accept the previously read EULA?
accept/decline/quit: accept
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 410.48?
(y)es/(n)o/(q)uit: yes
Do you want to install the OpenGL libraries?
(y)es/(n)o/(q)uit [ default is yes ]: yes
Do you want to run nvidia-xconfig?
This will update the system X configuration file so that the NVIDIA X driver
is used. The pre-existing X configuration file will be backed up.
This option should not be used on systems that require a custom
X configuration, such as systems with multiple GPU vendors.
(y)es/(n)o/(q)uit [ default is no ]: no
Note: I think there should be "no" and I donot know why :-)
Install the CUDA 10.0 Toolkit?
(y)es/(n)o/(q)uit: yes
Enter Toolkit Location
[ default is /usr/local/cuda-10.0 ]: /usr/local/cuda-10.0
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: yes
Install the CUDA 10.0 Samples?
(y)es/(n)o/(q)uit: yes
Enter CUDA Samples Location
[ default is /home/lab225 ]: /home/lab225/Documents/Samples/CUDA-10.0_Samples
sudo service lightdm start # Start graphical interfaceIf installation is successful, your should see the following output:
===========
= Summary =
===========
Driver: Installed
Toolkit: Installed in /usr/local/cuda-10.0
Samples: Installed in /home/lab225/Documents/Samples/CUDA-10.0_Samples
Please make sure that
- PATH includes /usr/local/cuda-10.0/bin
- LD_LIBRARY_PATH includes /usr/local/cuda-10.0/lib64, or, add /usr/local/cuda-10.0/lib64 to /etc/ld.so.conf and run ldconfig as root
To uninstall the CUDA Toolkit, run the uninstall script in /usr/local/cuda-10.0/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall
Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-10.0/doc/pdf for detailed information on setting up CUDA.
Logfile is /tmp/cuda_install_8756.logTo configure the CUDA environment for all users (and applications) on your
system create two files (use sudo and a text editor of your choice)
# Create file cuda.sh
sudo touch /etc/profile.d/cuda.sh
# Open cuda.sh file
sudo nano /etc/profile.d/cuda.sh
# Add content to the cuda.sh file
export PATH=$PATH:/usr/local/cuda/bin
export CUDADIR=/usr/local/cuda
# Also create file cuda.conf
sudo touch /etc/ld.so.conf.d/cuda.conf
# Open cuda.conf file
sudo nano /etc/ld.so.conf.d/cuda.conf
# Add content to the file
/usr/local/cuda/lib64
# Restart dynamic linker run-time bindings
sudo ldconfig
# Create symbolic links to GCC6 in the CUDA bin folder.
sudo ln -s /usr/bin/gcc-6 /usr/local/cuda-10.0/bin/gcc
sudo ln -s /usr/bin/g++-6 /usr/local/cuda-10.0/bin/g++
# Test CUDA by building the examples
# Copy the CUDA samples source directory to someplace in your home directory
# Go to the directory with the samples and run:
make -j4 # build all examples using 4 CPU processes
# There could be compilation error for the samples
# Error: cannot find -lGL
# Fix it by following the instructions in this link (final 2 commands):
# http://techtidings.blogspot.com/2012/01/problem-with-libglso-on-64-bit-ubuntu.html
sudo rm /usr/lib/x86_64-linux-gnu/libGL.so
sudo ln -s /usr/lib/libGL.so.1 /usr/lib/x86_64-linux-gnu/libGL.so
# Verify CUDA version
nvcc --version # 10.0 or higher
cat /usr/local/cuda/version.txt