Compute Cluster

• Term: Funded in 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 563020064

The purpose of this application is to replace or expand the central computing cluster operated by Leibniz Universität IT Services (LUIS) for Leibniz Universität Hannover (LUH) and to expand the GPU resources. As part of the “Scientific Computing” service, LUIS operates a medium-sized computer system to cover the basic need for scalable computing power and to process and store scientific simulation data. Over 800 scientists from more than 50 institutes use this computing cluster to investigate issues in their respective fields. As part of the centralization of services at Leibniz Universität Hannover, the LUIS offers the so-called “Research Cluster Housing” (FCH) in addition to the “Scientific Computing” service, which enables LUH institutes to integrate their own computing resources, procured in coordination with the LUIS, into the central computing cluster and thus use the existing infrastructure without significant administrative effort. Conversely, resources that are not yet fully utilized are made available to all cluster users, which not only leads to an improved overall utilization of the investment, but also indirectly to more computing power being available to the institute in the event of higher future demand. In addition to research, the computing cluster is also used for teaching purposes, currently mainly in the fields of artificial intelligence and machine learning. The “Scientific Computing” group provides a dedicated platform to support this. This gives both students and researchers access to the necessary computing resources for practical learning and experimentation in these pioneering disciplines. The computing cluster currently consists of almost 250 computing nodes (excluding the FCH partitions), all of which access a central parallel file system (“Lustre”) via a broadband InfiniBand interface. Despite the high demand from the user community, there is only a maintenance contract for the GPU resources and the MPP sub-cluster “Amo”. Many computing nodes need to be replaced with newer, more energy-efficient ones due to their age and increased maintenance costs. In addition, the computing cluster must be expanded to meet the ever-increasing demands of LUH scientists on computing capacity.

• DFG Programme: Major Research Instrumentation
• Major Instrumentation: Compute Cluster
• Instrumentation Group: 7000 Datenverarbeitungsanlagen, zentrale Rechenanlagen
• Applicant Institution: Gottfried Wilhelm Leibniz Universität Hannover
• Leader: Marcus Jabben