Parallel performance of an iterative solver based on the golub-kahan bidiagonalization
Kruse C, Sosonkina M, Arioli M, Tardieu N, Rüde U (2020)
Publication Type: Conference contribution
Publication year: 2020
Journal
Publisher: Springer
Book Volume: 12043 LNCS
Pages Range: 104-116
Conference Proceedings Title: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Event location: Bialystok
ISBN: 9783030432287
DOI: 10.1007/978-3-030-43229-4_10
Abstract
We present an iterative method based on a generalization of the Golub-Kahan bidiagonalization for solving indefinite matrices with a 2×2 block structure. We focus in particular on our recent implementation of the algorithm using the parallel numerical library PETSc. Since the algorithm is a nested solver, we investigate different choices for parallel inner solvers and show its strong scalability for two Stokes test problems. The algorithm is found to be scalable for large sparse problems.
Authors with CRIS profile
Involved external institutions
Old Dominion University
United States (USA) (US)
Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique (CERFACS)
France (FR)
Libera Università Mediterranea (LUM)
Italy (IT)
École Polytechnique - Université Paris-Saclay
France (FR)
United States (USA) (US)
Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique (CERFACS)
France (FR)
Libera Università Mediterranea (LUM)
Italy (IT)
École Polytechnique - Université Paris-Saclay
France (FR)
How to cite
APA:
Kruse, C., Sosonkina, M., Arioli, M., Tardieu, N., & Rüde, U. (2020). Parallel performance of an iterative solver based on the golub-kahan bidiagonalization. In Roman Wyrzykowski, Konrad Karczewski, Ewa Deelman, Jack Dongarra (Eds.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 104-116). Bialystok, PL: Springer.
MLA:
Kruse, Carola, et al. "Parallel performance of an iterative solver based on the golub-kahan bidiagonalization." Proceedings of the 13th International Conference on Parallel Processing and Applied Mathematics, PPAM 2019, Bialystok Ed. Roman Wyrzykowski, Konrad Karczewski, Ewa Deelman, Jack Dongarra, Springer, 2020. 104-116.
BibTeX: Download