UFZ-Rechenzentrum. Photo: Sebastian Wiedling / UFZ

EVE Cluster

The UFZ runs a High Performance Computing (HPC) cluster so big and complex calculations of the scientists can run locally at the UFZ. For example in different projects big satellite images will be analyzed to observe the developments of the climate.

Two prominent projects which use the cluster to run their calculations are the Drought Monitor Germany and the Forest condition monitor.

Many different languages and models are used on the cluster. On the one hand the go-to development languages on HPC cluster - Fortran and Python - and on the other hand at the UFZ self developed models like mHm. With the installed GPUs AI workflows can also run on the cluster.

Technical facts about the cluster:

  • 100 servers
  • 4500 CPU cores
  • 70 Terabyte RAM
  • 4 Petabyte storage
  • 15 Nvidia A100 GPUs
  • 100 Gigabit OmniPath Interconnect