This SGER award will support the parallel implementation of the immersed boundary method in the application of a three- dimensional heart model. The immersed boundary method is a computational technique for the simulation of systems in which a viscous incompressible fluid interacts with an immersed elastic boundary Such systems are ubiquitous in biological fluid dynamics, and the immersed boundary method has already been applied to problems of blood flow in the cardiac chambers, platelet aggregation during blood clotting, aquatic animal locomotion, and wave propagation in the inner ear. Current implementation of the immersed boundary method are fully vectorized on Cray computers, but only a single processor is used although several are available. The aim of the proposed research is to create a parallel implementation which maintains full vectorization while making effective use of all available processors. This work will be done on the Cray C-90 Computer, a 16-processor machine which is scheduled to arrive at the Pittsburgh Supercomputing Center this fall. The scientific utility of the immerse boundary method will be enormously enhanced by the anticipated speedup resulting from parallelization.
