Documentation
Research
HiPerGator

HiPerGator

This guide provides an overview of the UF’s high-performance cluster, HiPerGator. HiPerGator includes cutting-edge processors and NVIDIA’s Ampere A100 GPUs, making UF the first university globally to use this technology. Additionally, HiPerGator offers nodes optimized for memory-intensive tasks. Maintained by UFIT Research Computing, the cluster allows researchers to focus on their work without worrying about hardware or software maintenance.

What is and Why Use an HPC?

A High-Performance Computing (HPC) cluster is a collection of powerful servers managed by scheduling software like Slurm, which efficiently queues and runs user scripts. Here are some general reasons to use an HPC:

  • Enhanced Processing Power: HPC clusters offer significantly more computational power than standard personal computers, enabling faster data processing and analysis.

  • Scalability: HPC systems can handle large-scale computations and simulations, making them ideal for research that requires processing vast amounts of data.

  • Resource Efficiency: By utilizing shared resources, HPC clusters allow multiple users to perform intensive computations simultaneously, optimizing resource usage and efficiency.

HiPerGator Infrastructure

HiPerGator 3.0

  • Cores: 30,720 AMD EPYC 7702 Rome (2.0 GHz) and 9,600 AMD EPYC 75F3 Milan (3.0 GHz)
  • Memory: 8GB RAM per core
  • Storage: 4 Petabytes (PB) of Blue fast storage
  • GPUs: 608 NVIDIA RTX 2080TI and RTX 6000

HiPerGator AI NVIDIA DGX A100 SuperPod

  • Nodes: 140 NVIDIA DGX A100
  • Cores: 17,920 AMD Rome
  • GPUs: 1,120 NVIDIA Ampere A100
  • Storage: 2.5 PB All-Flash
  • Performance: 17.2 Petaflops (double precision), 0.7 Exaflops (AI operations)