Parallel sparse Gaussian elimination with partial pivoting

In this article, we consider the factorization of a sparse nonsymmetric matrix using Gaussian elimination with partial pivoting on a multiprocessor having a globally-shared memory. The parallel algorithm makes use of a static data structure developed by George, Liu and Ng in [17]. Some numerical experiments on a Sequent Balance 8000 are presented to demonstrate the efficiency of the parallel implementation.Research supported in part by the Applied Mathematical Sciences Research Program, Office of Energy Research, U.S. Department of Energy under contract DE-AC05-84OR21400 with Martin Marietta Energy Systems, Inc.