HPC Computing FAQs

  • August 2022, a DUO authentication step is now part of logging into the HPC.  
    For SSH, you will need to choose:
      1. Duo push to your phone number/app
      2. A phone call to your phone number
    For your SFTP client, you will also need to complete a DUO step.
  • Faculty can use the  to request accounts for themselves, their student assistants and for non-KSU collaborators. 
  • This usually means that you are not connected through the VPN properly. Make sure that the Global Protect shows as connected and that the vpn-groups.kennesaw.edu portal is selected.
  • Please contact Tom Boyle (x6180 or tboyle@kennesaw.edu) to discuss the needs of you and your collaborators and schedule a training session.

  • Once you have a connection through the VPN (vpn-groups portal), you will need to use an SSH program. Mac users can make use of the Terminal.app. Windows users can use Putty or Windows PowerShell.
    'ssh NetID_username@hpc.kennesaw.edu'

    For those wanting to utilize an XWindows session:

    As of 08/22, there is a DUO authentication step required for SSH. (Example) 

  • A list of the software installed on the cluster is maintained on the .
  • HPC users can request software using the .
  • Example: To have MATLAB load into your environment whenever you start a new session, use $ module initadd MATLAB.
    1. Use a graphical file transfer application that supports SFTP (via SSH)
      Example: Download Cyberduck for Mac or Win at .

    2. From within a local terminal, use the scp command.
      Example: scp C:\localfile.txt NetID@hpc:/data/home/your_NetID/

  • There are different file formats that can be 'unzipped':

    • gunzip will extract the contents of .gz files.
    • unzip will extract the contents of .zip files.
    • tar -xvf will extract the contents of .tar.gz and .tar.bz2 files.
  • If connected via SSH, use nano.
    $ nano <yourfilename>
    If you are consented via SFTP, you can right-click a file and select Edit to use an editor on your local machine.
  • Yes, the cluster currently has two nodes, each with 4 P100 NVIDIA cards. A graphical Processing Unit (GPU) is hardware specifically designed to perform graphics related tasks efficiently. Some software can be made to use GPUs for general purpose computing.

  • The KSU cluster has three major queues:

    • batch - this is the default queue and if comprised of only CPU nodes
    • himem - this queue has one CPU node with 1 TB of RAM
    • gpu - this queue has nodes with GPU capabilities
  • From the command line:
    qstat - Show status of PBS batch jobs
    showq - Displays information about active, eligible, blocked and completed jobs.
  • While there is no limit on the number of jobs, each user is limited by the total number of cores they can use concurrently. (480 cores, under heavy utilization 320 cores). Any jobs beyond the limit will be queued to run as soon as your other jobs complete and resources are available.
  • Walltime is the actual time an application runs, as in timer for the clock on the wall. The walltime is a parameter that a user provides in their job submission script.
  • This can happen when transferring files from system to system and the end of line character is now what is expected for Linux. A simple thing to try:
    $ dos2unix name_of_your_file
  • Please email the HPC support team at hpc@kennesaw.edu and Cc: service@kennesaw.edu.

vHPC FAQs

Pilot testing for the new HPC (vhpc) began in July of 2025.  The system should be available beginning some time in Fall semester 2025.

Place holder for launch information.

  • The vHPC consists of 12 nodes each with 2 CPUs, 4 GPUs and 512 GB of RAM.  Use the button below for more details.

    For referencing KSU facilities, use the persistent link to the maintained at the Library's DigitalCommons site.

    vHPC Facilities Statement
  • If you have a billable account set up for your project to use the vHPC, you can r for yourself and your team.
  • Yes.  Users should connect to the KSU VPN using the vpn-groups portal.
  • UITS maintains a wiki of technical documentation for the vHPC at
  • SLURM is the Simple Linux Utility for Resource Management.  It provides the resource manager and job scheduler functions.
  • For the latest list of software available, use the support the cluster: 

  • A KSU user is limited to using 144 CPU cores and 12 GPUs simultaneously.

    A single job is limited to a 720 hour walltime.

    A job on a single node is limited to 503 GB or RAM.

  • Contact tboyle@kennesaw.edu
  • Yes.  Learn more about this and other research computing cores.
  • User home directories are limited to 25 GB.  The staging directory can be used 
  • No.  To perform computation on  data requiring research compliance, other arrangements will need to be made.