Compute-Clustr

• Subject Area: Basic Research in Biology and Medicine
• Term: Funded in 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 439143547

In the newly founded "Center of Protein Assemblies" (CPA) at the Technical University of Munich, a compute cluster shall be purchased aused by several groups as a central element within the framework of a common IT usage concept. The compute cluster will be used by theoretically working groups (AG Zacharias, AG Antes) for the application of molecular simulations as well as experimental groups for image processing of microscopy images (AG Bausch) but in part also for evaluation of single molecule measurements, virtual screening and structural calculations ( groups in the CPA). The device is operated in close connection to the Leibnitz Rechenzentrum (LRZ), whereby a fast fiber optic data line allows the highly efficient exchange of data. It contains a 750 Tbyte file server system for further processing and reduction of extensive data. The computer codes of the compute cluster should consist of GPUs (Graphical Processing Units), which complement the complementary predominantly CPU-based computation / simulation options on the LRZ. The main method used by the theoretically working groups is molecular dynamics methods for which modern GPU systems (consumer cards) achieve the greatest efficiency and the best price / performance ratio in the majority of applications. But also for applications of exp. groups (e.g., image processing) exist high-efficiency GPU methods. As an accessory and integral part of the Compute cluster, cooling racks are also requested. The required cold storage space with an efficient outdoor water cooling system is provided for in the CPA (already under construction). In addition to the proximity of the groups of the CPA, the sharing of the Compute Cluster will contribute to the creation and implementation of new collaborations in the CPA.

• DFG Programme: Major Research Instrumentation
• Major Instrumentation: Compute-Cluster mit Datenserver
• Instrumentation Group: 7030 Dedizierte, dezentrale Rechenanlagen, Prozeßrechner
• Applicant Institution: Technische Universität München (TUM)
• Leader: Professor Dr. Martin Zacharias