A_SPARSE is an extremely reliable and efficient Iterative Solver Toolkit for solving large linear systems of the form AX=B where A is a sparse unsymmetric structured coefficient matrix. Such problems will typically arise from finite difference or finite volume approximations of a system of Partial Differential Equations on a topologically parallelepiped domain such as are found in 3-D Oil Reservoir Simulations, Semiconductor Device Simulations, and Computational Fluid Dynamics problems. A_SPARSE is based on newly designed imcomplete block triangular preconditioning strategies and stabilized variable block versions of GMRES, without restarts. These methods enable A_SPARSE to solve highly ill-conditioned linear systems having virtually any customer specified 3-D approximation stencil.
Russell Sabora
President
Elegant Mathematics, Inc.
12142 NE 166th Place
Bothell, WA 98011
USA
206-488-2061
206-488-7395 (fax)
solvers@elegant-math.com
For applications in related solution areas, see the following indices: Algebraic Computation, Industrial R&D, Mathematics, Numerical Analysis, Operations Research, Science and Mathematics, the developer index for Elegant Mathematics, Inc. and the market segment index for Math, Physics, Other Sciences.