The most powerful supercomputers in Europe have been offered to European researchers in academia and industry, as the PRACE Research Infrastructure (RI) launched the third call for access to its resources. In this call, a total of three Tier-0 supercomputers at the highest performance level are available for applicants, plus 17 national Tier-1 systems, via a synchronised DECI call. This is the first time the European Tier-1 resources have been available for use through a PRACE call. They were previously provided by DEISA (Distributed European Infrastructure for Supercomputing Applications) in DECI calls, now part of PRACE.
One year ago, at the end of a successful EC-funded preparatory phase project, PRACE (The Partnership for Advanced Computing in Europe) was created as one of the first Research Infrastructures listed on the ESFRI roadmap and started to provide high-end (Tier-0) supercomputer resources and services to the European scientific communities. Since then, it has managed to increase the aggregated capability of its Tier-0 systems by more than a factor of three. In addition to the access to Europe’s fastest supercomputers, PRACE also offers support for the preparation and optimisation of applications. This important element of the PRACE offer is supported through EC-funded implementation phase projects.
The latest Tier-0 system, ‘Hermit’ is available for researchers in this call for the first time. Hermit is a Cray XE6 system that is due to be installed in autumn 2011 at the Gauss Centre for Supercomputing (GCS) partner HLRS (High Performance Center of University Stuttgart) and has a current peak performance of one Pflop. This first installation step is to be followed by a four to five Pflops second step in 2013.
The one Pflop IBM BlueGene/P system, Jugene, hosted by the GCS partner Jülich Supercomputing Centre (JSC), Germany and the Bull Bullx cluster, Curie, hosted by GENCI partner CEA at Bruyères-Le-Châtel, France are also available for European researchers in the third PRACE call. Curie will reach a peak performance of 1.6 Pflops in its second installation phase in October 2011. The phase one system has been available for PRACE since January 2011.
The following Tier-1 systems are provided by the PRACE RI:
Cray XT4/5/6 and Cray XE6 – three large Cray XE and XT systems at EPCC (UK), KTH (Sweden) and CSC (Finland)
IBM Blue Gene/P – three BG/P systems at IDRIS (France), RZG (Germany) and NCSA (Bulgaria)
IBM Power 6 – three IBM Power 6 systems at RZG (Germany), SARA (The Netherlands) and CINECA (Italy)
Eight clusters at FZJ (Germany, Bull Nehalem cluster), LRZ (Germany, Xeon cluster), HLRS (Germany, NEC Nehalem cluster plus GP/GPU cluster), CINES (France, SGI ICE 8200), BSC (Spain, IBM PowerPC), CINECA (Italy, Westmere plus GP/GPU cluster), PSNC (Poland, Bullx plus GP/GPU cluster and HP cluster) and ICHEC (Ireland, SGI ICE 8200)
The third call closes on 22 June 2011, to be followed by other regular calls. There is also a continuous preparatory access call for the PRACE RI Tier-0 resources. The PRACE RI will be boosted by an IBM three Pflops Tier-0 system called ‘SuperMUC’ by the GCS partner LRZ (Leibniz Supercomputing Centre in Munich), available in mid-2012. Further Tier-0 installments to the PRACE RI will be procured in Italy and Spain.
