Setup new machine

setup_new_machine.sh

# Check number of cpus
grep -c ^processor /proc/cpuinfo     

sudo apt-get install tmux 
sudo apt-get install htop
sudo apt install git

# Conda
sudo mkdir -p /scratch/shubham/packages
sudo chown -R shubham /scratch/shubham
cd /scratch/shubham/packages
wget https://repo.anaconda.com/archive/Anaconda3-2019.10-Linux-x86_64.sh


# If you install ufw for firewall and security, dont forget to do; reference:
# https://www.digitalocean.com/community/questions/after-installing-the-firewall-can-not-log-in-with-ssh
sudo ufw allow ssh

# Get specs
lsb_release -a
cat /etc/os-release
hostnamectl
uname -r


# Change root password
sudo -i
sudo passwd

# Change user password
sudo passwd

# add user
sudo adduser XYZ

# Customize your prompt
# Ref: https://www.ostechnix.com/hide-modify-usernamelocalhost-part-terminal/
echo "export PS1="\[\033[01;32m\]\u@gpu:\w\$ "" >> ~/.bashrc

# Reboot remote machine
# Ref: https://phoenixnap.com/kb/restart-linux-using-command-prompt
reboot
# sudo shutdown -r now

# Copy ssh keys
# Ref: https://apple.stackexchange.com/questions/210109/how-to-store-ssh-credentials-in-terminal
mkdir -p ~/.ssh
cat ~/.ssh/id_rsa.pub | ssh ${machine_name} "cat >> ~/.ssh/authorized_keys"

# CuDNN
# https://stackoverflow.com/a/36978616
sudo apt install nvidia-cuda-toolkit

# Check cuda version
# https://medium.com/@changrongko/nv-how-to-check-cuda-and-cudnn-version-e05aa21daf6c
cat /usr/local/cuda/version.txt

# If still having issues with cuda for RNN in torch: RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED for rnn
conda install pytorch==1.0.1 torchvision==0.2.2 cudatoolkit=10.0 -c pytorch
# Might run into numpy errors then https://github.com/pytorch/pytorch/issues/2731
pip install -U numpy

# Verify using https://stackoverflow.com/a/48152675