Two new high-performance computers make Rhineland-Palatinate a leader in scientific computing
Super-fast computing capacity will support and strengthen top-flight research in Mainz and Kaiserslautern
The two newly installed high-performance mainframe computers at Johannes Gutenberg University Mainz and at the University of Kaiserslautern provide the German federal state of Rhineland-Palatinate a new national profile in high-performance computing. The state government, the German Research Foundation (DFG), and both universities invested a total of almost EUR 5.5 million in the new high-performance computers, thus pushing the state of Rhineland-Palatinate to the forefront of research in this area. In addition, the now available super-fast computing capacity will support and strengthen top-flight research in Mainz and Kaiserslautern, regardless of whether it be in meteorology, genome research, chemistry, physics, mechanical engineering, or in new materials research.
"The integration of both mainframe computers in the Alliance for High-Performance Computing Rhineland-Palatinate is setting new standards in the field of high-performance computing for the German federal state of Rhineland-Palatinate as a location of excellent universities and research. What is particularly positive is the fact that scientists all over the state will profit from it," says Minister of Science Doris Ahnen. "Computer-based research has enormous potential and anybody who recognizes this and wants to benefit from it can now find optimum conditions for keeping up with the latest developments in the area. The technical requirements are now in place for top-quality international research across all scientific disciplines." Furthermore, as Minister Doris Ahnen added, the state not only financed the building of both supercomputers but has also focused on creating on a good technical basis and server structure. Thus, a large share of the funds from Germany's Economic Stimulus Package (Konjunkturprogramm II) was invested in information technology. Of the total EUR 86 million, approximately EUR 10.6 million alone went into IT equipment.
Both systems are available not only to researchers at the universities in Mainz and Kaiserslautern but are also integrated in the Alliance for High-Performance Computing Rhineland-Palatinate (AHRP). The AHRP bundles the computing capacity of the state to make it available for all universities and research facilities in Rhineland-Palatinate. It also offers courses (lectures, seminars, internships) and consulting for those using the high-performance computer. The start-up of the two new high-performance computers and the activities of the AHRP thus make an essential contribution to promoting all of the universities in Rhineland-Palatinate and improving their international competitiveness.
The new cluster offers scientists at Johannes Gutenberg University Mainz the opportunity to use high-performance computers with state-of-the-art technology. "Many scientific disciplines now use high-performance computers to form models for complex systems. Numeric simulations are thus developing into a third pillar of scientific research along with the classic pillars of theory and experiment," explains the President of Johannes Gutenberg University Mainz, Professor Dr. Georg Krausch. "Since the complexity of interconnections in fields such as meteorology, genome research, or the interaction of individual molecules cannot always be depicted in the design of experiments, modern scientific research is increasingly using simulations. And this is no different at our university. The further advancement of this area is one of the core aims of our JGU Research Unit Computational Sciences in Mainz. Therefore, we are thrilled to have the high-performance cluster as support for our research unit on the one hand, and, on the other hand, the high-performance computer will also help maintain the outstanding positions of our working groups in the natural sciences at the forefront of their research and further expand those positions."
"High-performance computing has long been an important building block for research in physics, chemistry, biology, and mechanical engineering at the University of Kaiserslautern," says its President, Professor Dr. Helmut J. Schmidt. "Disciplines such as the social sciences, economics, and computer science are also using more and more high-performance computing for the simulation of complex and highly parallel systems and to better understand the base correlations in them. With the new high-performance computer in Kaiserslautern, our researchers here now have the right tool to advance their research. The utilization of such systems stands and falls with a suitable support structure, which the University of Kaiserslautern with its corresponding course and seminar offerings already has in place."
Users of the AHRP in Rhineland-Palatinate will benefit from the fastest connection between two German universities at speeds of 120 GBit/s. Researchers can select which of the two high-performance computers is most suitable for their purposes and have the simulation data exchanged immediately. Simply because it would be impossible to send the quantity of data produced over the public Internet, a close connection of the data processing computer centers is required.
The high-performance cluster at the Data Center of Johannes Gutenberg University Mainz can carry out 287,000 billion computations per second. "The mainframe computer uses a mere 14 cabinets, making it very compact for its capacity. To build a comparable computer from PCs would require 5,000 individual computers and take up twenty times as much space," explains Professor Dr. André Brinkmann, head of the Data Center at JGU. The advancements in computer technology, along with those in cooling technology, make it possible for the new cluster to work at a full capacity of 480 kW with twice the energy efficiency of previous clusters. The name of the cluster is MOGON – from the Roman name Mogontiacum, the Latin name for the city of Mainz.
The new high-performance cluster will mainly be used by institutes and working groups from the field of physics, but also researchers in mathematics, biology, medicine, and the geosciences increasingly take advantage of the opportunities afforded by the high-performance computer. At JGU, complicated questions for computer technology also come from the discipline of molecular genetics. In order to decode the genetic information of living organisms as completely as possible using next-generation sequencing of DNA molecules, genome researchers need enormous quantities of computing and storage capacity. Questions from biomedicine and biology examined using the computer come from research in oncology and evolution.
Each and every tumor has a unique genetic makeup. For patient-specific cancer therapy, it is vital to understand which genes specify a tumor and whether these genes can be made use of in therapy. Understanding the consequences of climate change can be fundamentally improved if we better understand how important parts of the food chain, such as insect larvae living in water, react to changes in temperature as the climate fluctuates. The simultaneous determination of the activities of all genes through mass sequencing helps provide information in this area. With the help of Illumina sequencing equipment, molecular genetics can encode several billion base pairs of multiple human genomes in two weeks. In doing so, each genome requires a terabyte of storage. Due to the prevailing technical conditions, each genome, i.e., the total of genetic information, has to be sequenced into sequence snippets of 100 building blocks (base pairs). Using these billions of snippets, scientists then use bioinformatics to reconstruct the entire DNA sequence of chromosomes. All this uses massive amounts of hard disk memory and requires more than 512 GB RAM per computer.
The high-performance computer at the University of Kaiserslautern is run by the Regional University Computing Center Kaiserslautern. "A separate room with a modern water cooling system and a separate power supply was created for the new cluster," explains Professor Dr. Paul Müller, head of the Computing Center. "The first expansion stage of the cluster is to be installed in seven cabinets and the room has still enough space for the second expansion phase planned for 2013. In addition to the performance of the cluster, the project also focused considerably on energy efficiency and the efficient use of space." The Kaiserslautern cluster was named "Elwetritsch" as homage to a mythical creature of southwest Germany.
High-performance computing at the University of Kaiserslautern is used mainly by engineering and natural sciences working groups and institutes. Mechanical engineering researchers use it to perform flow simulations, simulation of the settlement characteristics of bulk solids, and thermodynamic calculations. In physics, thousands of differential equations need to be solved quickly to simulate, for example, the behavior of ferromagnets subjected to short-term laser impulses. Chemists at the University of Kaiserslautern use it, for example, to investigate how to avoid chemical waste products in synthesizing processes or to carry out calculations to better understand the electronic structure of atoms and molecules. In biology, high-performance computing is important for gene sequence analyses or for phylogenetics, which deals with the development and relationship between living beings. To meet these diverse challenges, the high-performance computer in Kaiserslautern is equipped with various hardware components that offer the optimum environment for each respective area of application.