A block Recycled GMRES method with investigations into aspects of solver performance

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Parks, Michael L., Soodhalter, Kirk M., Szyld, Daniel B.
Format: Preprint
Published: 2016
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910213372641280
author Parks, Michael L.
Soodhalter, Kirk M.
Szyld, Daniel B.
author_facet Parks, Michael L.
Soodhalter, Kirk M.
Szyld, Daniel B.
contents We propose a block Krylov subspace version of the GCRO-DR method proposed in [Parks et al.; SISC 2005], which is an iterative method allowing for the efficient minimization of the the residual over an augmented Krylov subspace. We offer a clean derivation of our proposed method and discuss methods of selecting recycling subspaces at restart as well as implementation decisions in the context of high-performance computing. Numerical experiments are split into those demonstrating convergence properties and those demonstrating the data movement and cache efficiencies of the dominant operations of the method, measured using processor monitoring code from Intel.
format Preprint
id arxiv_https___arxiv_org_abs_1604_01713
institution arXiv
publishDate 2016
record_format arxiv
spellingShingle A block Recycled GMRES method with investigations into aspects of solver performance
Parks, Michael L.
Soodhalter, Kirk M.
Szyld, Daniel B.
Numerical Analysis
65F10
We propose a block Krylov subspace version of the GCRO-DR method proposed in [Parks et al.; SISC 2005], which is an iterative method allowing for the efficient minimization of the the residual over an augmented Krylov subspace. We offer a clean derivation of our proposed method and discuss methods of selecting recycling subspaces at restart as well as implementation decisions in the context of high-performance computing. Numerical experiments are split into those demonstrating convergence properties and those demonstrating the data movement and cache efficiencies of the dominant operations of the method, measured using processor monitoring code from Intel.
title A block Recycled GMRES method with investigations into aspects of solver performance
topic Numerical Analysis
65F10
url https://arxiv.org/abs/1604.01713