server_setup.md

Server Setup

Created: Mar 31, 2023 Updated: Nov 28, 2023 (by Zilong Wang)

0. SSH-Key

ssh-copy-id -i ~/.ssh/id_rsa.pub YOUR_USER_NAME@IP_ADDRESS_OF_THE_SERVER

1. oh-my-zsh

1. Install zsh+omz

   • https://github.com/ohmyzsh/ohmyzsh/wiki/Installing-ZSH

   • https://ohmyz.sh/#install

     # Install Zsh
     sudo apt install zsh
     sh -c "$(wget https://raw.githubusercontent.com/ohmyzsh/ohmyzsh/master/tools/install.sh -O -)"

2. Themes

   • Powerlevel10k: https://github.com/romkatv/powerlevel10k
   • 
   • Install:

     git clone --depth=1 https://github.com/romkatv/powerlevel10k.git ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/themes/powerlevel10k

   • Set ZSH_THEME="powerlevel10k/powerlevel10k" in .zshrc
   • Solve bugs: add unset ZSH_AUTOSUGGEST_USE_ASYNC at the bottom of .zshrc

3. Plugins

   • git
   • zsh-autosuggestions:
     • git clone https://github.com/zsh-users/zsh-autosuggestions ${ZSH_CUSTOM:-~/.oh-my-zsh custom}/plugins/zsh-autosuggestions
   • zsh-syntax-highlighting:
     • git clone https://github.com/zsh-users/zsh-syntax-highlighting.git ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/plugins/zsh-syntax-highlighting
   • In .zshrc:

2. Conda

1. Recommend to use miniconda and use a separate env other than base for development (so that if you somehow ruin your env in use, you can simply reinstall it.)

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh

2. Create Environment

conda create -n general python=3.X

3. PyTorch

1. Web Page: https://pytorch.org/get-started/locally/

2. Install a specific cuda toolkit

• Option 1: manually install the cuda toolkit

  • Google search for the runfile of the right version
  • sudo sh cuda_XXX_linux.run
  • Double check the symlink in /usr/local/cuda
  • Add this to the .zshrc or .bashrc

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

• Option 2: install cuda toolkit as a package in conda env

  • All you need is simply nvcc
  • https://anaconda.org/nvidia/cuda-nvcc
    • conda install -c nvidia cuda-nvcc
    • conda install -c "nvidia/label/cuda-1X.X.X" cuda-nvcc
  • Make sure there is no code to import/specify the cuda directory in your .bashrc or .zshrc

3. Check nvcc version: nvcc -V
4. Why do my ncvv -v and nvida-smi show different versions?
   • nvidia-smi shows the highest cuda version the current cuda env supports. (could be higher than your actual cuda version).
   • nvcc -v is your true cuda version.

3. Tmux

.tmux.conf:

wget https://raw.githubusercontent.com/Shang-Data-Lab/CodeBase/main/code/tmux.conf
mv tmux.conf .tmux.conf

4. Vim

.vimrc: https://github.com/amix/vimrc

git clone --depth=1 https://github.com/amix/vimrc.git ~/.vim_runtime
sh ~/.vim_runtime/install_basic_vimrc.sh

echo "set cursorline" >> .vimrc