High-Performance-Computing Cluster

• Term: Funded in 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 547936783

In recent decades, Halle has developed into one of the technology centers in Central Germany. The Martin Luther University Halle-Wittenberg is significantly involved in this development. The close integration of theoretical and experimental research is characteristic of scientific and medical research areas. However, the existing computing capacity is no longer sufficient for the necessary data analyzes and simulations. A weak point in the existing infrastructure is primarily in the area of computing massively parallel jobs, such as those used in modeling biological effects, in virtual screening, in simulating large networks, in sequence and expression data analysis ("deep sequencing") or in advanced use of artificial intelligence ("deep learning"). In the current high-performance cluster "Janus" at the Martin Luther University Halle-Wittenberg, in addition to the lack of computing power, wear and tear is also becoming a growing problem. After 8 years of use, the failure rate of the computing technology has increased significantly. Due to the outdated hardware, repairs are only possible with remanufactured spare parts. These are becoming increasingly difficult to obtain and do not have the life expectancy of new parts. The hardware no longer meets current requirements in terms of both the required computing power and energy efficiency. The HPC software is largely outdated and can no longer be updated due to expired licenses. This requires complex manual adjustments to current scientific program packages. In individual cases, even this is no longer possible, so decentralized HPC solutions have to be created. The increasing number of failures and the lack of computing power are having a noticeable negative impact on study operations and research activities. Due to hardware repairs and necessary software adjustments, the time required to maintain the cluster is constantly increasing. It is therefore necessary to replace the existing high-performance cluster "Janus" with a more modern, more powerful and at the same time more energy-efficient system that will also meet the demands of the coming years.

• DFG Programme: Major Research Instrumentation
• Major Instrumentation: High-Performance-Computing Cluster
• Instrumentation Group: 7000 Datenverarbeitungsanlagen, zentrale Rechenanlagen
• Applicant Institution: Martin-Luther-Universität Halle-Wittenberg
• Leader: Sebastian Lux