MPI+OpenMP parallel implementation of explicitly preconditioned conjugate gradient method
Abstract:
A preconditioner for large sparse symmetric positive definite coefficient matrix is considered based on its approximate inverse in the form of product of a lower triangular sparse matrix by its transpose. A parallel algorithm for the construction and application of the preconditioner is proposed with the use of MPI+OpenMP techniques. Comparative timing results for the MPI+OpenMP and MPI implementations of the considered preconditioning and the Jacobi preconditioning used with the conjugate gradient method for a model problem and the University of Florida collection test problems is presented.
Keywords:
iterative solution of linear systems, sparse matrices, incomplete inverse triangular factorization, parallel preconditioning, conjugate gradient method
