High Performance Computing
High-performance computing (HPC) is a cornerstone of data-, model-, and simulation-driven research: For many questions in the complex, co-evolving dynamics of human-Earth system research to climate science, scientific progress depends on access to reliable and modern HPC resources. At the same time, the HPC landscape is evolving rapidly and hardware architectures, software ecosystems, and best practices for performance and reproducibility are continuously changing.
For MPI-GEA, the HPC Core Unit is a dedicate core facility that provides tailored HPC support for the specific scientific needs of researchers at the institute. Services offered by the core unit include:
- Operation of the institute-owned cluster GEANY (together with the MPCDF, see below)
- Software development with a focus on HPC environments
- Trainings and workshops for researchers, ‘concierge’ service for HPC-related tasks
- Individual consultations for researchers who wish to use HPC resources
- Networking with other HPC facilities
- Close collaboration with all other departments and facilities of the institute, to optimise data flows and to enable FAIR research
Our cluster GEANY
The HPC Core Unit operates the institute-owned cluster GEANY. As of January 2026, GEANY provides researchers at the institute with the following resources:
- 79 nodes in total
- 19,584 logical CPUs
- 62 TB RAM
- 4.7 PB GPFS-based storage
- 20 NVIDIA GPUs/accelerators (16x NVIDIA L40S, 4x NVIDIA H100 SXM)
- 200 GB/s Mellanox Technologies InfiniBand fabric for node interconnect
GEANY provides researchers with a modern software stack for HPC-based research and queueing/resource management based on Slurm.