|
Sibirskii Zhurnal Vychislitel'noi Matematiki, 2012, Volume 15, Number 2, Pages 223–228
(Mi sjvm474)
|
|
|
|
Iterative solver for systems of linear equations with a sparse stiffness matrix for clusters
A. A. Kalinkinab, Yu. M. Laevskyab a Institute of Computational Mathematics and Mathematical Geophysics (Computing Center), Siberian Branch of the Russian Academy of Sciences, Novosibirsk
b Novosibirsk State University, Novosibirsk
Abstract:
In this paper, a package of programs for solving systems of linear equations with a sparse matrix for computers with distributed memory is proposed. This package is based on the iterative algorithm for solving the initial system of equations with preconditioner constructed using the algebraic domain decomposition. Such an approach makes possible to multiply by the preconditioner and a stiffness matrix on cluster. Also, to improve the efficiency of computation, PARDISO and SparseBlas functionalities from Intel$\circledR$MKL library are used on each process. In addition to parallelization among processes, this package uses OpenMP parallelization on each of these processes as well as Intel$\circledR$MKL internal functional parallelization.
Key words:
sparse solver, domain decomposition, parallelization, MPI and OpenMP.
Received: 06.10.2011 Revised: 17.01.2012
Citation:
A. A. Kalinkin, Yu. M. Laevsky, “Iterative solver for systems of linear equations with a sparse stiffness matrix for clusters”, Sib. Zh. Vychisl. Mat., 15:2 (2012), 223–228; Num. Anal. Appl., 5:2 (2012), 182–186
Linking options:
https://www.mathnet.ru/eng/sjvm474 https://www.mathnet.ru/eng/sjvm/v15/i2/p223
|
Statistics & downloads: |
Abstract page: | 748 | Full-text PDF : | 159 | References: | 42 | First page: | 18 |
|