Pegasus — AI High Performance Computing @ DFKI #

This page contains the user documentation for our HPC cluster Pegasus.

Pegasus facts #

• Hosted @ DFKI Kaiserslautern
• 100+ PFLOPs for AI workloads
• 300+ GPUs
• 5500+ CPU Cores
• 40+ TiB RAM
• Slurm Workload Manager
• Enroot container environments
• Custom monitoring suite

Document structure #

For information on how to access our infrastructure, terms of use, etc. consult our guidelines.

For the technicalities on how to make your code work and get the most out of our hardware, see the section about our Slurm cluster.

For questions not yet answered, check the Questions / Contact page.