Scalable Parallel Algorithms for Unstructured Problems

Abstract

In this paper we summarize our work on development of parallel algorithms for searching large unstructured trees, and for finding solution of large sparse systems of linear equations. Search of large unstructured trees is at the core of many important algorithms for solving discrete optimization problems. Solution of large sparse systems of equations is required for solving many important scientific computing problems. For both of these domains, we show that highly scalable parallel algorithms can be developed, and these algorithms can obtain high speedup on a large number of processors.