Reaching the scalability of a distributed AMG solver

View at publisher
Authors: Igor Konshin, Kirill M. Terekhov
Cited by: 1
References: 26

Abstract

Abstract In this paper, a distributed parallel version of algebraic multigrid method is proposed for linear algebraic systems arising from discretization of systems of scalar elliptic equations. The implementation demonstrates good scalability on thousands of cores and surpasses the open source packages HYPRE BoomerAMG and PETSc GAMG in performance on a number of systems. Effective scalable implementations of algorithms in the setup stage of the algebraic multigrid method are considered. The possibility of reducing the number of processors in communicator for coarse systems has been implemented. Variants of cycles of the multigrid method are considered, which make it possible to obtain independence of the number of iterations as the system size increases without losing the scalability of the method. These developments form the basis for further extension of the method for multiphysical problems.

Fields of Study & Subjects

Advanced Numerical Methods in Computational Mathematics Matrix Theory and Algorithms Aquatic and Environmental Studies

Publication details

Journal / source
Russian Journal of Numerical Analysis and Mathematical Modelling
Publisher
Walter de Gruyter GmbH
Volume
41
Issue
1
Pages
45–67
Publication date
1 February 2026
Language
EN