DATA-PARALLEL SPARSE LU FACTORIZATION

Sparse matrix factorization is a computational bottleneck in many scie ntific and engineering problems. This paper examines the problem of fa ctoring large sparse matrices on data-parallel computers. A multifront al approach is presented in which only the fine-grain concurrency foun d within the elimination of each supernode is exploited. Throughput ap proaching that of large dense matrix factorizations is demonstrated on two data-parallel systems, the MasPar MP-2 and the Thinking Machines CM-5.