The aim of this paper is to apply the factorized sparse approximate inverse (FSAI) preconditioner to the iterative solution of linear systems with indefinite symmetric matrices. Until now the FSAI technique has been applied mainly to positive definite systems and with a limited success for the indefinite case. Here, it is demonstrated that the sparsity pattern for the preconditioner can be chosen in such a way that it guarantees the existence of the factorization. The proposed scheme shows excellent parallel scalability, performance and robustness. It is applicable with short recurrence iterative methods such as MinRes and SymmLQ. The properties are demonstrated on linear systems arising from mixed finite element discretizations in linear elasticity. Copyright © 2009 John Wiley & Sons, Ltd.
miércoles, 23 de diciembre de 2009
Factorized parallel preconditioner for the saddle point problem
The aim of this paper is to apply the factorized sparse approximate inverse (FSAI) preconditioner to the iterative solution of linear systems with indefinite symmetric matrices. Until now the FSAI technique has been applied mainly to positive definite systems and with a limited success for the indefinite case. Here, it is demonstrated that the sparsity pattern for the preconditioner can be chosen in such a way that it guarantees the existence of the factorization. The proposed scheme shows excellent parallel scalability, performance and robustness. It is applicable with short recurrence iterative methods such as MinRes and SymmLQ. The properties are demonstrated on linear systems arising from mixed finite element discretizations in linear elasticity. Copyright © 2009 John Wiley & Sons, Ltd.
Suscribirse a:
Enviar comentarios (Atom)
No hay comentarios:
Publicar un comentario
Nota: solo los miembros de este blog pueden publicar comentarios.