Skip to content

FF chevron

Getting help using the HPC cluster

You can get assistance for using the cluster by contacting the HPC systems staff. The best way to reach us is by opening a ticket. In some situations, you may wish to have a phone/zoom conversation as well; in that case, we would still like you to initiate the process by opening a ticket so that we can coordinate time and method.

For new users, users with limited programming experience, and those who want to refresh their knowledge of HPC related topics we recommend to attend our introductory workshops.

Before you open a ticket ...

  • Please use the search box at top right corner on this site
  • See our FAQ
  • Do a Google search for any error messages
  • Use an AI engine such as ChatGPT or Gemini

How to open a ticket

Use any of the following ways open a ticket in the DIT ticket system. We highly recommend you first read the tips for submitting a ticket below.

Contact the DIT Service Desk

Phone: (301)405-1500
Web: https://itsupport.umd.edu/
E-mail: itsupport@umd.edu

Contact HPC systems staff directly

E-mail: hpcc-help@umd.edu

Tips for submitting a ticket

When opening a ticket, please include a good description of the issue you've encountered.

When updating an already open ticket, it is best to reply back to the email you received from the ticket system. Using the ticket system's web portal is less helpful in adding details sometimes required in technical questions.

  • Always open a new ticket for a new issue. (ie: brand new email)
  • Do include the following info:
    • Job ID
    • Filename and directory location of the job submit script
    • Directory location where we can find log and output files
    • Description of the problem (in your own words)
  • Do NOT include these:
    • Do not include passwords/passphrases
    • Do not attach files unless we request
    • Do not take screenshots of text output
    • Do not copy/paste very long output/error text
  • Use the script command to show us the interactive CLI sessions
  • Do let us know when a proposed solution works as this helps us build a better knowledge base