NOTE: Previously there were separate desktop options for OpenGL and CUDA sessions.  These are now combined and provide access to 1 GPU, 4 CPUs, and 40GB of RAM.  In order to share the visualization servers evenly, users are unable to change these settings.  If your program requires more resources, please contact CCR Help to discuss possible alternatives.

OpenGL and CUDA Desktop 

This app launches a CUDA/OpenGL-enabled session on the remote visualization nodes.  If you do not require hardware accelerated graphics, the UB-HPC or faculty cluster desktop sessions may start sooner than this and offer longer wall times. The longest allowable wall time for these nodes is 24 hours. However, you can select less if desired and this will potentially decrease your wait time.

To learn more about on how to Use Interactive Sessions in OnDemand please click here.

NOTE: Starting a visualization session can take anywhere from a few minutes (if the visualization nodes have available capacity) to up to 24 hours (if there are current jobs running and the nodes are full).  The visualization session you're starting is an actual job on the UB-HPC cluster and therefore subject to the same issues that may affect jobs submitted directly to the batch scheduler, SLURM.  The scheduler looks at all jobs submitted and puts them in priority order while checking to see what nodes (servers) are available for use. There are two visualization nodes that can be shared by twelve jobs each.  Users may only run one visualization session at a time.  Visualization sessions are permitted to run for 24 hours and will be canceled when time expires, whether you close the session or not.  Please make sure you're monitoring your sessions so you save your files before you're logged off.  PLEASE be fair to other users and log off when you're done rather than waiting for time to expire.

  1. Slurm Account: Your default Slurm account will be used if nothing else is entered.  To see what Slurm accounts you have access to, use the 'slimits' command in a terminal window.
  2. Number of hours: The longest amount of time you can request is 24 hours.
  3. Exclude host: If you want to ensure your job doesn't run on a specific node, enter the node short name here.  This is in the form of: cpn-rack#-u# i.e. cpn-u22-01  We recommend you only do this if told to by a CCR staff member. 

In this example, we've started up an Viz Node - OpenGL/CUDA Desktop.  Once you click the Launch button, you will be redirected to the session information where you can see the status of the job.  This one is queued:

If you have elected to receive an email when your job starts, you can leave the CCR OnDemand portal until the email arrives and then log back in to open your session.  

As it is preparing to start you will see:

When a session is ready to launch, it will look like this in the Interactive Sessions section of CCR OnDemand:

Click on the blue "Launch Viz Node - OpenGL/CUDA Desktop" button and you will see your desktop window on the node you've been allocated:

You are now logged into a visualization session and onto a node in the CCR academic cluster.  You can begin your work!  

If this is your first time starting a remote visualization session, you may be prompted with this pop-up:

Please click "Use default config"

NOTE: If your session starts and does not display a menu or icons on the bottom (as shown in these screenshots), check out this troubleshooting article

To start a terminal window, click on the window icon at the bottom of your screen, shown here:

This is a terminal window and where you will type all of your commands.  If you are unfamiliar with UNIX commands, we recommend you view the following articles:

Basic UNIX

UNIX File Transfer

To learn more about launching GUIs for COMSOL, MATLAB, StarCCM, and ABAQUS click here.

For information about where files are stored, how to check your quotas, and how to purchase additional storage, please see:

Storage and Quotas

All software available for use on CCR servers is listed here:

Current Software Information

How do I install software?

Still unsure what to do next?  Schedule an appointment with one of our computational scientists for a training session by submitting a help request.

To learn how to delete or end a running session click here.