Efficiency of general Krylov methods on GPUs - An experimental study

Anzt H, Dongarra J, Kreutzer M, Wellein G, Köhler M (2016)

Publication Status: Published

Publication Type: Conference contribution, Conference Contribution

Publication year: 2016

Publisher: IEEE Computer Society

Pages Range: 683-691

Article Number: 7529929

ISBN: 9781509021406

DOI: 10.1109/IPDPSW.2016.45

Abstract

This paper compares different Krylov methods based on short recurrences with respect to their efficiency whenimplemented on GPUs. The comparison includes BiCGSTAB, CGS, QMR, and IDR using different shadow space dimensions. These methods are known for their good convergencecharacteristics. For a large set of test matrices taken from theUniversity of Florida Matrix Collection, we evaluate the methods'performance against different target metrics: convergence, number of sparse matrix-vector multiplications, and executiontime. We also analyze whether the methods are «orthogonal»in terms of problem suitability. We propose best practicesfor choosing methods in a «black box» scenario, where noinformation about the optimal solver is available.

Authors with CRIS profile

Moritz Kreutzer Regionales Rechenzentrum Erlangen (RRZE) Gerhard Wellein Professur für Höchstleistungsrechnen

Related research project(s)

Equipping Sparse Solvers for Exascale II (ESSEX-II) (SPPEXA) SPP 1648: Software for Exascale Computing Jan. 1, 2016 - Dec. 31, 2018

Involved external institutions

Oak Ridge National Laboratory US United States (USA) (US) Max-Planck-Institut für Dynamik komplexer technischer Systeme / Max Planck Institute for Dynamics of Complex Technical Systems DE Germany (DE)

How to cite

APA:

Anzt, H., Dongarra, J., Kreutzer, M., Wellein, G., & Köhler, M. (2016). Efficiency of general Krylov methods on GPUs - An experimental study. In Proceedings of the 30th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2016 (pp. 683-691). IEEE Computer Society.

MLA:

Anzt, Hartwig, et al. "Efficiency of general Krylov methods on GPUs - An experimental study." Proceedings of the 30th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2016 IEEE Computer Society, 2016. 683-691.

BibTeX: Download