DEISA – Meeting the European High Performance Computing Challenge

The provision of high performance computing resources to researchers has traditionally been the objective and mission of a few national HPC centres in Europe. Increasing global competition between Europe, USA, and Japan and growing demands for compute resources at the highest performance levels on one hand, and stagnant if not shrinking budgets combined with the need to innovate faster on the other hand, begin to show the limits of this concept. To stay competitive major investments are needed every two years – an innovation cycle that even the most prosperous countries hesitate to fund.

To advance science in Europe eight leading European HPC centres devised an innovative strategy to build a Distributed European Infrastructure for Scientific Applications (DEISA). The centres partner in building and operating a tera-scale supercomputing facility. This becomes possible through deep integration of existing national high-end platforms, tightly coupled by a dedicated network and supported by innovative system and grid software. The resulting virtual distributed supercomputer has the capability for natural growth in all dimensions without singular procurements at the European level. The investments that DEISA partners undertake in the course of their development naturally increase the capacity of the overall DEISA infrastructure. Advances in network technology and the resulting increase in bandwidth and lower latency virtually shrink the distance between the nodes in the distributed super-cluster. Furthermore, DEISA can expand horizontally by adding new systems, new architectures, and new partners thus increasing the capabilities and attractiveness of the infrastructure in a non-disruptive way.

Figure 1: Map of Europe with logos of partners

Most important for the success of DEISA is the advancement of science in Europe. DEISA collaborates with leading European research groups initially from the following scientific and industrial disciplines:

• Material Sciences
• Cosmology
• Fusion Research
• Life Sciences
• Coupled CFD and CAA Applications
• Environmental Sciences.

The researchers aim to demonstrate that DEISA will enable new research results and enhanced scientific output in yet unprecedented ways. The list of disciplines listed above is not exhaustive, and new scientific initiatives are being prepared.

Status of DEISA

Led by IDRIS-CNRS (France) the DEISA project started its activities in May 2004 with seven partners: FZJ and RZG in Germany, CINECA in Italy, EPCC and ECMWF in the UK, CSC in Finland, and SARA in the Netherlands. The DEISA infrastructure is being implemented in two phases. Initially, four “core” partners have coupled their systems using virtually dedicated 1 Gbit/s connections provided by the pan-European research network GÉANT and the national research networks RENATER (France), DFN (Germany), and GARR (Italy). Within one year other systems will be integrated into the DEISA super-cluster. At end of 2004 is combines over 4000 IBM Power 4 processors and 416 SGI processors for an aggregate peak performance of 22 teraflops.

The DEISA research infrastructure uses proven system software like AIX and Linux, GPFS and LoadLeveler, and grid middleware like UNICORE and Globus to achieve both a tight coupling of systems and access to heterogeneous resources.

DEISA will be open to collaboration with other Europe HPC centres and related initiatives world-wide. During its project span of five years it will adapt to the rapidly changing IT technology to retain its leadership position and guarantee its persistence beyond the project lifetime.

The big challenge is to demonstrate the DEISA strategy for the production of first class computational science as part of the European e-infrastructure.

For further information see the DEISA Consortium at www.deisa.org.

DEISA is funded in part by the European Commission in the 6th framework program under grant number 508630.

• Dietmar Erwin
Central Institute for Applied Mathematics (ZAM), Research Centre Jülich

• Victor Alessandrini
Institut du Développement et des Ressources en Informatique Scientifique (IDRIS)

  top