Title: Parallel Iterative Algorithms for Large Sparse Linear Systems

Speaker: Maxim Naumov, Purdue University

Date/Time: Monday, November 9, 2009,  9:00 AM       

Location: CSRI Building, Room 279 (Sandia NM)

Brief Abstract: The solution of linear systems is important in many areas of computational science and engineering. In particular, numerical handling of partial differential equations (PDEs) plays a crucial role in modeling of physical processes. It involves discretization of these PDEs using for example finite difference or finite element methods and often requires solution of large sparse linear systems. We focus on the parallel iterative algorithms for large sparse nonsymmetric linear systems. The most popular iterative schemes for these problems are part of Krylov subspace family of methods, and include BiCGStab, GMRES and CGNR methods. We look at their reliability using ILU/IQR-preconditioning techniques and suggest two alternative schemes. First is a hybrid scheme based on algebraic domain decomposition techniques that uses direct and iterative algorithms to solve in parallel a single linear system. Second is a novel iterative scheme that uses deflation-based preconditioned block-row projection method with an outer iterative solver to solve in parallel a single linear system.

CSRI POC: Mike Heroux, (320) 845-7695