LINDAR for Networks of Shared Memory Muiltprocessors

Period of Performance: 12/28/1993 - 12/27/1995

$500K

Phase 2 SBIR

Recipient Firm

Scientific Comput Assoc
One Century Tower, 265 Church Street
New Haven, CT 99999
Principal Investigator

Research Topics

Abstract

The Navy, DOD, and the civilian economy can be made more effective and efficient by the aggressive use of high performance computing. Clusters of commodity, RISC microprocessors, including shared-memory multiprocessor versions, offer the potential of very high aggregate performance at very low hardware costs. The catch is the cost and time required to develop the required parallel applications software for these clusters. We propose a novel solution to the parallel software crisis; the Trellis Visualization System which generalized and extends current scientific workbench technology, eg. Khoros and AVS, to handle general computing problems in the parallel cluster environment.