Project Details
Projekt
Parallel Cluster for Centralization
Subject Area
Computer Science
Term
Funded in 2025
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 553406827
Focus areas in simulation at the Technische Universität Bergakademie Freiberg include mathematical and engineering problems in the fields of finite element methods and the numerical description of structural and fluid mechanics processes, which are often coupled with mass and heat transport, as well as quantum chemical calculations and structural mechanical computations. Specifically, this involves the development and application of parallel algorithms in mathematics and computer science (simulation, data, artificial intelligence), theoretical physics (density functional theory and molecular dynamics, semiconductor research, calculation of IR and Raman vibrational spectra, etc.), geophysics, rock mechanics, and solid mechanics, as well as structural and fluid mechanics. Recently, artificial intelligence has gained significant importance. There is also a trend towards Scientific Machine Learning in Freiberg, meaning that simulation methods are increasingly combined with machine learning. Users at TU Freiberg have utilized both local and regional computing resources (NHR, Tier 0/1, international) for their computational needs; before 2019, significant computing resources were decentralized at the institutes. The compute cluster established in 2019 successfully replaced decentralized computing resources by centralizing computational needs. Indeed, professorships have used their funds to expand the central cluster: the 2019 device has been expanded five times (mechanics and fluid dynamics twice, numerical mathematics and optimization, stochastic, geophysics, and geoinformatics). The cluster acquired in 2019 is now reaching the end of its life cycle and is due for replacement. Moreover, in the area of artificial intelligence, a trend towards rebuilding decentralized (GPU) computing resources is emerging, as the 2019 cluster is not adequately equipped for this. The goal is to avoid decentralized solutions and offer an attractive, central solution. This avoids binding personnel at the institutes for operation and maintenance, strengthens IT security, and increases utilization efficiency through shared use. Further synergy effects arise from shared use of licenses, as well as central climate and energy technology. The device also facilitates the transition to other systems: a structured approach with preparatory calculations, scalability tests, and code optimizations can support the efficient use of higher-level HPC resources (e.g. NHR). The low-barrier access to scientific computing resources at the TU BAF's URZ has clearly broadened the user base.
DFG Programme
Major Research Instrumentation
Major Instrumentation
Parallel-Cluster zur Zentralisierung
Instrumentation Group
7000 Datenverarbeitungsanlagen, zentrale Rechenanlagen
Applicant Institution
Technische Universität Bergakademie Freiberg
