High Performance Computing in Science and Engineering '09

At the end of the year 2008, we have seen a strategic step towards a funct- ning HPC infrastructure on Tier-0 level in Germany. Based on an agreement ( Verwaltungsabkommen") between the Federal Ministry of Education and " Research (BMBF) and the state ministries for research of Baden-Wurttem- .. berg, Bayern, and Nordrhein-Westfalen, a budget of overall 400 Million Euro had been allocated - equally shared between federal and state authorities in a ?ve year time frame - to establish the next generation of HPC systems at the Gauss Centre for Supercomputing (GCS) - consisting of the three nat- nal supercomputing centres HLRS (Stuttgart), NIC/JSC (Julich), and LRZ .. (Munich). As part of that strategic initiative, in May 2009 already NIC/JSC has installed the ?rst phase of the GCS HPC Tier-0 resources, an IBM Blue Gene/P with roughly 300. 000 Cores, this time in Julic .. h, With that, the GCS provides the most powerfulhigh-performance computing infrastructure in - rope already today. HLRS and its partners in the GCS have agreed on a common strategy for the installation of the next generation of leading edge HPC systems.Over the next few years, HLRS and LRZ as the other two GCS centers will upgrade their systems accordingly. The plan is to have a Tier-0 HPC system within GCS operating at any time in this ?ve year period. Asanintermediatestep,HLRShasreplacedmostoftheirNECSX-8nodes by the NEC SX-9/12M192, a system with roughly 20 TFLOPs peak.