Install notes for private computers
GIGABYTE based BIOS setting
- First, DO NOT PLUG IN your GPU until the driver is set up.
- Internal Graphic : Auto > Enable
- Display Priority : PCIe > Internal
NVIDIA driver download .run file : click here
If you click download from the above site, you will get a .run file format for installing drivers.
Before you run the .run file, you first need to stop your Xserver display manager.
Press [Ctrl] + [Alt] + [F1], enter the script below
$ service --status-all | grep dm
(Result) [+] [:dm]
The part described as [:dm] is your display manager.
Substitute the [:dm] part below with the result of the script above.
$ sudo service [:dm] stop
(Result) * Stopping Light Display Manager [:dm]
Run the code below. Press 'Yes' for every option they ask.
$ sh <DIR where you downloaded the .run file>/NVIDIA-Linux_x86_64-390.87.run
In cases, the newest nvidia driver provided in the homepage happens to not work.
You may simply determine your latest available driver here, then install via PPA.
sudo apt-get purge nvidia* # Only if you have any previous versions of nvidia driver installed.
# add PPA
sudo add-apt-repository ppa:graphics-drivers
sudo apt-get update
# for GeForce 10x series
sudo apt-get install nvidia-390
# prevent unwanted automatic updates
sudo apt-mark hold nvidia-390
After you have successfully installed, you shall see the same results when typing the code below.
$ nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.87 Driver Version: 390.87 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 TITAN X (Pascal) Off | 0000:4B:00.0 Off | N/A |
| 67% 86C P2 249W / 250W | 5026MiB / 12221MiB | 82% Default |
+-------------------------------+----------------------+----------------------+
| 1 TITAN X (Pascal) Off | 0000:4C:00.0 On | N/A |
| 88% 90C P2 225W / 250W | 3842MiB / 12213MiB | 78% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
$ sudo reboot
You can skip the step above and automatically install the driver within CUDA installation.
Check the details below.
CUDA download page : click here
Before executing the file, stop the display manager by following the description above.
$ sudo sh <DIR where you downloaded the .run file>/cuda_9.0.176_384.81_linux.run
$ sudo apt-get install vim
$ git clone https://github.com/amix/vimrc.git ~/.vim_runtime
# Awesome version
$ sh ~/.vim_runtime/install_awesome_vimrc.sh
# Basic version
$ sh ~/.vim_runtime/install_basic_vimrc.sh
Open your ~/.bashrc file.
vi ~/.bashrc
# Press Shift + G, Add the lines on the bottom
export PATH=$PATH:/usr/local/cuda/bin
export LD_LIBRARY_PATH=/usr/local/cuda/lib64
export CUDA_HOME=/usr/local/cuda
To check if the CUDA toolkit is successfully installed, type the line below.
$ nvcc --version
* (Result)
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176
cuDNN download page : click here
(Membership is required, just sign in!)
Download the newest cuDNN v7.0.
$ cd <DOWNLOAD DIR>
$ tar -zxvf ./cudnn-9.0-linux-x64-v7_0_5.tar.gz
$ sudo cp cuda/include/cudnn.h /usr/local/cuda/include/
$ sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/
$ sudo chmod a+r /usr/local/cuda/include/cudnn.h
$ sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
$ sudo apt-get install vim
$ git clone https://github.com/amix/vimrc.git ~/.vim_runtime
# Awesome version
$ sh ~/.vim_runtime/install_awesome_vimrc.sh
# Basic version
$ sh ~/.vim_runtime/install_basic_vimrc.sh
- Disable vim folding
$ vi ~/.vimrc
set nofoldenable
Linux Mint 19 한글 입력 here
$ sudo apt-add-repository ppa:hodong/nimf
$ apt update
$ apt install nimf nimf-libhangul
$ im-config -n nimf
$ apt autoremove fcitx-*
Set as 'Hangul'
$ sudo vi /usr/share/X11/xkb/symbols/altwin
# Line 5
key <RALT> { type[Group1] = "TWO_LEVEL",
symbols[Group1] = [Alt_R, Meta_R] };
# Change the above lines into
key <RALT> { type[Group1] = "TWO_LEVEL",
symbols[Group1] = [ Hangul ] };
Install figlet
$ sudo apt-get install figlet
$ sudo vi /etc/bash.bashrc
# Press [Shift] + [G] and write the code on the bottom
clear
printf "Welcome to Ubuntu 18.04.5 LTS (GNU/Linux-Mint-19 x86_64)\n"
printf "This is the install guide made by Bumsoo Kim.\n\n"
printf " * Documentation: https://github.com/meliketoy/bumsoo.install\n\n"
printf "##############################################################\n"
figlet -f slant "Bumsoo Kim"
printf "\n\n"
printf " Data Mining & Information System Lab\n"
printf " GPU Computing machine : bumsoo@[:local_ip]\n\n"
printf " Administrator : Bumsoo Kim\n"
printf " Please read the document\n"
printf " https://github.com/meliketoy/bumsoo.install/README.md\n"
printf "##############################################################\n\n"
- netrc configuration
$ sudo vi ~/.netrc
machine github.com
login [:username]
password [:password]
- Git initial settings
$ git config --global user.name "[:user_name]"
$ git config --global user.email [:user_email]
- git alias settings
$ git config --global alias.last 'log -p -1 HEAD'
$ git config --global alias.lg 'log -10 --pretty=format:"%h - %an : %s" --graph' # [hash / name / msg]
For jupyter notebook configuration, type in the command line below.
$ jupyter notebook --generate-config
* Result :
Writing default config to: <HOME_DIR>/.jupyter/jupyter_notebook_config.py
$ vi ~/.jupyter/jupyter_notebook_config.py
presh [Esc], then enter /ip to find the ip configuration. You will find the line below
## The IP address the notebook server will listen on.
#c.NotebookApp.ip = 'localhost'
Erase the '#' and change it into ...
c.NotebookApp.ip = '163.152.163.112' # the ip address for your server
presh [Esc], then enter /port to find the port number. You will find the line below
## The port the notebook server will listen on.
#c.NotebookApp.port = 8888
Erase the '#' and enter whatever port number you want
c.NotebookApp.port = 9999
python -m ipykernel install --user --name [virtualEnv] --display-name "[displayKenrelName]"
# Export
conda env export > environment.yml
# Download
conda env create -f environment.yml
Now, Enjoy! :)