Setting up Quadcorn for remote use
posted on August 12, 2016

Here, we document how to configure quadcorn for remote access.

We mostly followed the instructions here.

On the server

Setup user accounts on quadcorn

>> sudo useradd -m --create-home USERNAME

Choose password:

>> sudo passwd USERNAME

See here if you forgot -m.

Once you have created the users, setup the usergroup.

>> sudo groupadd -r GROUPNAME

Add each user to the group.

>> sudo usermod -a -G GROUPNAME user1
>> sudo usermod -a -G GROUPNAME user2

Start the SSH service.

>> /etc/init.d/ssh start

Open the SSH Daemon config file.

>> atom /etc/ssh/sshd_config

Add the following lines to only allow our user group:

#Which groups of users to allow

Edit the port number with a custom PORT_NUMBER:

#What ports, IPs and protocols we listen for
#Port 22

Check that X11 forwarding is set to yes:

X11Forwarding yes

We used:


Once this is done, just restart the ssh server:

sudo service ssh restart

On your client

The IP address of our server will be given to you.

To login to the server, type:


But we can do better than remembering all these details. We need to let the ssh program know the host IP address, the user login, and the port to listen to. To do so, create the ~/.ssh/config file:

>> atom ~/.ssh/config`

Type in the following:

Host quadcorn
  ForwardX11 yes
  Compression yes
  Ciphers blowfish-cbc

Now we can simply do this instead:

>> ssh quadcorn

This replaces the much more cumbersome option where you need to remember everything.

Customize your settings to use global software installs

Once you login to quadcorn, to make sure that you can use most of the installed software, you need to edit your ~/.profile file.

>> atom ~/.profile

Now add the following lines:

# set PATH to include anaconda global directory

# set PATH, LD_LIBRARY_PATH for cuda
export PATH="/usr/local/cuda/bin:$PATH"
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64"

# to get the bash shell, otherwise you get `/bin/sh` by default

To test that the path to python is set correctly, type:

>> which python

You should see:


To test whether cuda is installed correctly, type:

>> nvcc -V

You should see something like:

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2015 NVIDIA Corporation
Built on Tue_Aug_11_14:27:32_CDT_2015
Cuda compilation tools, release 7.5, V7.5.17