Skip to content

HiPerGator

For classroom and teaching support on HiPerGator, see UF Research Computing's guide: it.ufl.edu — Teaching Support.

Getting Started: From Zero to a Running Notebook

1. Get a HiPerGator Account

docs.rc.ufl.edu/quickstart/zero_hipergator

  • Every group needs a sponsor (PI) with an existing HiPerGator account and a funded resource allocation
  • Group members then request their own account under that sponsor
  • Once the sponsor approves, account creation takes about 2–3 business days

2. Ways to Connect

docs.rc.ufl.edu/quickstart/interfaces

  • Command line — connect over SSH into a Linux shell; best for automation and full flexibility
  • Web interfaces — Open OnDemand (GUI apps, RStudio, MATLAB) and JupyterHub (notebooks), both in your browser

3. Launch a Notebook via JupyterHub

jupyterhub.rc.ufl.edu

  • Go to jupyterhub.rc.ufl.edu and log in with your GatorLink credentials
  • Pick a resource profile from the dropdown — CPU cores, memory, GPUs, and time
  • Click Start — your request is submitted to the SLURM scheduler and the server usually starts within a few minutes

4. Stop Your Server When Finished

  • From the notebook's File menu, choose Hub Control Panel
  • Click Stop My Server
  • This frees the CPU/GPU/memory for the rest of your group — don't leave idle servers running

JupyterHub's Hub Control Panel

5. Choosing a Kernel on HiPerGator

  • UFRC managed kernels (e.g. UFRC Python-3.x) give quick access to hundreds of preinstalled packages
  • For reproducible, project-specific work, build your own conda environment and register it as a personal kernel
  • Same idea as Creating & Exporting a Conda Environment — just running on HiPerGator instead of your own machine

Training

UF Research Computing's HiPerGator Training page has recorded sessions and self-paced materials covering:

  • Intro to Research Computing and HiPerGator, and Linux command line basics
  • Slurm job submission (including MPI jobs)
  • Open OnDemand, Jupyter notebooks, Conda environments, and VSCode remote tunnels
  • Data transfer (Globus, SFTP, rsync, rclone)
  • Git/GitHub basics (self-paced Canvas course)

Each topic links to both a studio-recorded video and a Zoom recording of the live session.