Parallelization Strategies for Element-by-Element PCG Solver using HPF for Unstructured FE Applications on Linux clusters

Parallel implementation of an unstructured Finite Element (FE) solver using the Preconditioned Conjugate Gradient (PCG) Method is described here. High Performance Fortran (HPF) has been used with the implementation based on a 32-node Pentium II 350MHz cluster running LINUX. The PCG solver is set up for the Element-by-Element (EBE) Method. While this is a highly suitable method for the solution of very large problems, it is not inherently parallelizable in a distributed memory environment. The di culties associated with the parallelization of this algorithm and two ways to overcome it, in this environment, are presented here. Initial results have been presented for fairly large problems ranging from 500,000 to 1,600,000 unknowns using an 8-noded isoparametric quadrilateral element.