Planned center for High Performance Computing (HPC) at the Institute of Computer Science will bring together various research projects and will provide an improved level of HPC services
The Carl Zeiss Foundation will be providing a total of EUR 750,000 over four years to fund the Competence Center for HPC in the Natural Sciences at the Institute of Computer Science of Johannes Gutenberg University Mainz (JGU). The proposed competence center for high performance computing (HPC), which will be headed by Professor Bertil Schmidt, General Manager of the Institute of Computer Science, and Professor André Brinkmann, Director of the JGU Center of Data Processing, will promote interdisciplinary collaborations between the natural sciences and computer science at Mainz University over the long term. "HPC plays an important role in the science-oriented fields of our university. The use of computer simulations is one of the most important techniques, in addition to modeling and experimentation, for generating new insights in the natural sciences. HPC has thus become a factor that enhances the profile of Mainz University and has contributed decisively to the competitiveness of our research," said Professor Bertil Schmidt. "The new competence center will add to the reputation of our university in the field of HPC – particularly in view of the planned acquisition of the new supercomputer MOGON II and the potential for collaboration with the Center for Computational Sciences Mainz (CSM). The new HPC competence center will contribute over the long term to enhancing JGU’s profile in the computer and natural sciences in the areas of simulation and the evaluation of Big Data."
Many branches of the natural sciences are currently in the process of transition to the use of data-driven concepts. The storage and analysis of the huge amounts of data routinely generated in biology, physics, meteorology, and other disciplines is increasingly causing problems for the natural sciences. In general, the only solution for this is to develop novel, scalable algorithms and software and make use of HPC. However, the mere presence of computing resources is not sufficient unless the necessary methodological skills are also available, not only within the natural sciences but also in the fields of algorithm and program development and their implementation on modern HPC computer architectures.
With regard to translational research, the objective of the new Competence Center for HPC in the Natural Sciences is to facilitate the successful transfer of research results in computer science (i.e. design, implementation, and evaluation of scalable methods for analyzing and storing large amounts of data) so that these can be employed within the natural sciences. "The new competence center will therefore focus on research in the areas of Big Data and HPC and at the same time specifically devote itself to interdisciplinary collaborations with the users," added Professor André Brinkmann. "To meet these goals, the nature of the center must be oriented towards both research and provision of services. Whereas the implementation, expansion, and maintenance of user-friendly programs will clearly be a service aspect, the design and optimization of the programs on modern HPC computer architectures will be associated to many interesting research problems."
In particular, the competence center will work on applications in the fields of bioinformatics, the analysis of large amounts of data from particle accelerators, the identification and localization of meteorological structures, and the geosciences. The center will be focusing on the areas of hardware accelerators, benchmarking and application optimization, data mining, visual analytics, and stochastic optimization. It is also planned to create suitable program libraries to provide for the widest possible reutilization of results.
Mainz-based researchers are worldwide leaders in the field of simulation-driven research and have demonstrated their capabilities through their achievements in the PRISMA Cluster of Excellence, the Graduate School of Excellence "Materials Science in Mainz" (MAINZ) as well as in various collaborative research centers.
In April 2014, the German Council of Science and Humanities approved JGU's application for funding of a new supercomputer, MOGON II, thus providing a further impetus towards the consistent further development of scientific computing in Rhineland-Palatinate. A total of EUR 8.7 million will be invested in the new supercomputer by the federal government, the state government, and JGU in the period 2015 to 2017 to ensure that the Rhineland-Palatinate researchers within the Alliance for High-Performance Computing Rhineland-Palatinate (AHRP) are provided with top-class computing power until 2019. In addition to the needs of the researchers, the German Council of Science and Humanities also took into account the fact that the necessary methodological and operational expertise in the area of HPC is already available at Mainz University. These aspects are currently being expanded within the JGU Center of Data Processing and the Center for Computational Sciences Mainz.