Max Planck Society awards developers Cactus

Potsdam, 08 December 98 An ongoing project which culminated in a demonstration of the potentials of long-distance visual supercomputing and networking at SC98 won the Heinz-Billing-Preis, an award given each year by the Max Planck Society for an outstanding work in computational science. The project involves an international team of researchers that looks at new ways to use supercomputing to study the 3D Einstein equations and other problems of astrophysics. The group also developed Cactus. a code developed specifically to study problems such as colliding black holes, neutron stars, the formation of singularities, and other aspects of Einstein's theory that cannot be handled by analytic means.

Among the achievements of the research group are the development of Cactus, The demonstration that resulted from this collaborative effort showed the results of computations done simultaneously on supercomputers in Germany and the United States. The resulting 3D simulation of colliding neutron stars was displayed on the SC98 conference floor in Orlando.

Cactus was originally designed and created as a collaborative tool by Joan Masso and Paul Walker, former researchers at the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign. It was then jointly developed by researchers at AEI, Washington University in St. Louis and NCSA.

An NCSA team, led by Senior Research Programmer John Shalf, also developed the networking and VR software used in the visual supercomputing project. Cactus is being further developed by the same collaborators into a computational toolkit for solving range of equations on parallel machines. The toolkit will be made available to the world-wide user community early in 1999.

In addition, the project makes use of Globus, an integrated set of software components for high-performance networks developed by Ian Foster at Argonne National Laboratory and Carl Kesselman at the University of Southern California's Information Sciences Institute. The long-distance, real-time supercomputing, an essential component of the research project, involved using MPICH-G, a new Globus-enabled implementation of the Message Passing Interface (MPI).

The research group also depended on the STAR TAP interconnection point in Chicago to link information run on remote supercomputers to the National Science Foundation's very high-performance Backbone Network Service (vBNS). Jason Novotony and Meghan Thornton of the Distributed Applications Support Team, which is part of National Laboratory for Applied Network Research (NLANR), along with Warren Smith of Argonne, researchers from AEI, the Konrad-Zuse-Institut in Berlin, and the RZG in Garching, Germany, also worked with the team.

The US National Computational Science Alliance is a partnership to prototype an advanced computational infrastructure for the 21st century and includes more than 50 academic, government and industry research partners. The Alliance is one of two partnerships funded by the Science Foundation's Partnerships for Advanced Computational Infrastructure (PACI) programme, and receives cost-sharing at partner institutions. NSF also supports the Partnership for Advanced Computational Infrastructure (NPACI), led by the San Diego Supercomputer Center.

Sandra Wermer