NONLINEAR STRUCTURAL RESPONSE USING ADAPTIVE DYNAMIC RELAXATION ON A MASSIVELY PARALLEL-PROCESSING SYSTEM

A parallel adaptive dynamic relaxation (ADR) algorithm has been develo ped for non-linear structural analysis. This algorithm has minimal mem ory requirements, is easily parallelizable and scalable to many proces sors, and is generally very reliable and efficient for highly non-line ar problems. Performance evaluations on single-processor computers hav e shown that the ADR algorithm is reliable and highly vectorizable, an d that it is competitive with direct solution methods for the highly n on-linear problems considered. The present algorithm is implemented on the 512-processor Intel Touchstone DELTA system at Caltech, and it is designed to minimize the extent and frequency of interprocessor commu nication. The algorithm has been used to solve for the non-linear stat ic response of two- and three-dimensional hyperelastic systems involvi ng contact. Impressive relative speed-ups have been achieved and demon strate the high scalability of the ADR algorithm. For the class of pro blems addressed, the ADR algorithm represents a very promising approac h for parallel-vector processing.