40th TOP500 List: GCS Systems Placed 1st and 2nd on European Ranking, Claiming Positions 5 and 6 World-Wide
PRESS RELEASE 09/2012:
JUQUEEN First European System to Pass 5 Petaflops Barrier
Berlin/Germany, November 28, 2012 – The German Gauss Centre for Supercomputing (GCS) continues to be a key player in the world of High Performance Computing (HPC). According to the latest TOP500 list (released Nov. 12, 2012), GCS systems JUQUEEN and SuperMUC are the two most powerful supercomputers in Europe. With a peak performance of 5.033 Petaflops, GCS system JUQUEEN, installed at Jülich Supercomputing Centre (JSC), is the first HPC system in Europe to pass the 5 Petaflops barrier (quadrillions of calculations per second). It holds 5th place on the world-wide listing, while GCS system SuperMUC, installed at Leibniz Supercomputing Centre (LRZ) in Garching/Munich takes position 6 with a peak performance of 3.185 Petaflops. System Hermit, installed at GCS centre HLRS in Stuttgart, continues to be the world’s fastest supercomputer used for industrial development, research and science. With a peak performance of more than 1 Petaflops, Hermit marks position 7 on the European count.
While the Linpack performance of the three GCS systems is deciding criteria for the ranking on the renowned TOP500 list, the directors of the GCS supercomputing centres lay special emphasis on their systems’ energy efficiency and sustained performance. "JUQUEEN is targeted to tackle comprehensive and complex scientific questions, called Grand Challenges", explains Prof. Thomas Lippert, Director of JSC. "Projects from various scientific areas can profit from the system’s performance, e.g. in the areas of neuroscience, computational biology or energy and climate research. It enables complicated calculations in quantum physics, which were not possible before.“ JSC’s new Queen of European Supercomputers, an IBM BlueGene/Q system based on the IBM POWER architecture hosting 393,216 compute cores, is especially designed for compute intensive, highly scalable applications which can run in parallel on a very high number of compute cores. The system is amongst the most energy-efficient supercomputers in the world. With a performance/power ratio of approximately 2 Gigaflop/s per Watt, JUQUEEN provides a fivefold improvement in energy efficiency compared to its predecessor JUGENE thanks to a direct water cooling system that takes away heat directly from the processors.
The same is true for SuperMUC, installed at LRZ. The Garching supercomputer, an IBM System X iDataPlex consisting of 155,000 cores, which was officially inaugurated in July 2012, is also „not only fast but highly energy-efficient“. Its peak performance of about 3 Petaflops places the system second on the list of Europe’s fastest supercomputers to date while grabbing position 6 on the current TOP500 world-wide ranking. Prof. Arndt Bode, Director of LRZ, stresses the extreme energy efficiency of the system which is achieved through hot-water cooling technology. “Our PUE* value of 1.1 is a ratio currently unmatched by any supercomputer of comparable performance,” affirms Prof. Bode proudly. The system, designed as general purpose HPC system, allows an exceptionally versatile deployment: “SuperMUC is extremely user friendly! We run more than 150 different applications on our system per year, thus an instruction set allowing easy adaption of user software was a core requirement on the system architecture,” adds Prof. Bode. SuperMUC is being used for a wide spectrum of science and research tasks, ranging from medical and engineering & energy applications to astrophysics.
* PUE is the ratio of total amount of power used by a computer data centre facility to the power delivered to computing equipment.
With a peak performance of over 1 Petaflops, HLRS’s flagship computer Hermit continues to hold the title as the world’s fastest supercomputer used for industrial research (TOP500, Nov. 2012, sub-list Industry). Like the HPC systems of the other two GCS supercomputing centres, the Cray XE6 system is highly energy-efficient. Officially commissioned in February 2012, Hermit is the only system amongst its international competition dimensioned not only for science and research activities but also for use in industrial production processes. “When configuring Hermit, we concentrated on the fact that the enormous computing performance can be made available to science as well as business. Special emphasis was therefore placed on the sustained performance for real applications“, explains Professor Michael M. Resch, Director of HLRS. The computing power of Hermit is for example used by a team under Prof. Dr.-Ing. Ewald Krämer, Institute of Aerodynamics and Gas Dynamics of University Stuttgart. A total of 72 million core hours were allocated to their project Laminar Turbulent Flow Simulation, while Prof. Dr.-Ing. Wolfgang Schröder of RWTH Aachen (Fluid Mechanics and Institute of Aerodynamics) was granted 72.7 million core hours on Hermit to pursue research on project Direct Numerical Simulation of the Flow in an Internal Combustion Engine.
The world’s fastest supercomputer to date is Titan, a Cray XK7 system installed at Oak Ridge National Laboratory (USA) with 560,640 processors, including 261,632 NVIDIA K20x accelerator cores. Delivering a peak performance of 20.13 Petaflops, Titan knocked Lawrence Livermore National Laboratory’s Sequoia (17.59 Petaflops) out of No. 1 and into second place.
GCS Systems for European Science and Research Activities
All three GCS HPC systems – JUQUEEN, SuperMUC and Hermit – qualify as „Tier-0“ system in the European research infrastructure offered through the Partnership for Advanced Computing in Europe (PRACE). As a Hosting Member of PRACE, the Gauss Centre for Supercomputing offers in total more than 9 Petaflops of computing power for large-scale scientific projects to scientists and researchers from Europe and beyond. Computing time will be allocated to applicants based on a single peer review process. In the latest PRACE Regular Call for Proposals (October 2012), record-breaking project allocations of computing core hours were allocated to all three GCS systems:
JUQUEEN: Heavy ion phenomenology form lattice simulations by Szabolcs Borsanyi of the
Bergische Universität Wuppertal (Germany) received 91,791,360 core hours on JUQUEEN
SuperMUC: Physics-based evaluation of seismic shaking in Northern Italy (SHAKEIT) by Andrea Morelli of the Istituto Nazionale di Geofisica e Vulcanologia (Italy) received 53.4 million core hours.
Hermit: Large scale molecular dynamics simulations of nucleation by Jürg Diemand of the University of Zürich (Switzerland) received 35.3 million core hours on Hermit.
For further information on JUQUEEN, please refer to http://www.fz-juelich.de/ias/jsc/juqueen
For further information on SuperMUC, please refer to http://www.lrz.de/supermuc
For further information on Hermit, please refer to http://www.hlrs.de/hermit
The Gauss Centre for Supercomputing (GCS) combines the three national supercomputing centres HLRS (High Performance Computing Center Stuttgart), JSC (Jülich Supercomputing Centre), and LRZ (Leibniz Supercomputing Centre, Garching) into Germany’s Tier-0 Supercomputing institution. Concertedly, the three centres provide the largest and most powerful supercomputer infrastructure in Europe to serve a wide range of industrial and research activities in various disciplines. They also provide top-class training and education for the national as well as the European High Performance Computing (HPC) community. GCS is the German member of PRACE (Partnership for Advance Computing in Europe), an international non-profit association consisting of 25 member countries, whose representative organizations create a pan-European supercomputing infrastructure, providing access to computing and data management resources and services for largescale scientific and engineering applications at the highest performance level.
GCS has its headquarters in Berlin/Germany.
Regina Weigand, GCS Public Relations
+49 711 685-87261
This press release as pdf: