Littleton, Massachusetts-based Alliant Computer Systems Corp reports that Michigan Technological University, Argonne National Laboratory, and the Centre for Supercomputing Research & Development at the University of Illinois are to combine their Campus/800 80860 RISC-based systems to create a 200-processor massively parallel system for tackling extremely large problems: the combined system will have peak performance of 8 GFLOPS, and the three systems will be linked by a high-speed network; applications range from pollution modelling to development of parallel software; shared memory support is implemented at the ClusterNode level, with each node consisting of a cluster of 25 RISC processors and 4Gb of high-speed shared memory in the Campus 800; distributed memory support is provided via standard techniques such as message-passing, which enables communication between the ClusterNodes; with Co-operative Supercomputing, the ClusterNodes can be physically distributed across several organisations and geographical sites and connected over wide area T1, T3 or Sonet networks.
