Cluster user interfaces

User interaction with the cluster can take place in several ways:

  • via command line (remote terminal);
  • via graphical interface of application software (GUI);
  • via web interface.

Command line

This is a traditional manner of accessing the cluster. The user connects via the Internet to a remote terminal (usually with the SSH protocol), where you can run commands in text mode. You can edit and compile code, queue tasks (simulations), and monitor progress.

You can also call graphical tools from the command line.

Graphical interface of application software

When you connect to a cluster with an SSH connection (for example, with program PuTTY), you can use not only the terminal command line, but also graphic tools. SSH provides graphical session forwarding (X11 forwarding) from cluster to the user’s computer. 

Use the –X parameter when you connect to a cluster in a Linux environment.

ssh -X [username]

In order to use this function for MS Windows environment, additional preparatory steps must be made on the user’s PC:

  1. Install the tool X Windows server that you can download from here: 
  2. Run it in the background as a service. The first time you start, you may be prompted to change the settings, leave the default ones.
  3. In addition, you’ll need a PuTTY configuration for X Windows. Change the saved configuration for the cluster access server by making the following changes:

Keep the PuTTY configuration!

An example: Calling MATLAB software from the command line

To start the Matlab graphical interface on the cluster access server (, connect to the cluster command line by using the graphic session transfer described before.  

On the access server command line, run:

module load matlab/R2017b 

A program window will open on your computer, but all actions will run remotely on the cluster. The MATLAB interface provides the necessary tools for queuing and executing a task on computing nodes. 

Application software graphical interface

Some software packages (such as Schrodinger, MATLAB, etc.) have built-in support for parallel computing on the cluster. The user can choose to run the simulation locally without leaving the usual software environment, or queue it for execution on the cluster. In application software such as Schrodinger, working with the cluster can open a graphical interface both locally on your personal computer and remotely on the cluster control node. 

Web interface

Access to the cluster using a HTTP protocol or by opening a graphical interface (portal) on an Internet browser. See the example in the image below — Moab Viewpoint interface.