Stable iterative refinement algorithms for solving linear systems

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wu, Chai Wah, Squillante, Mark S., Kalantzis, Vasileios, Horesh, Lior
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913907504840704
author Wu, Chai Wah
Squillante, Mark S.
Kalantzis, Vasileios
Horesh, Lior
author_facet Wu, Chai Wah
Squillante, Mark S.
Kalantzis, Vasileios
Horesh, Lior
contents Iterative refinement (IR) is a popular scheme for solving a linear system of equations based on gradually improving the accuracy of an initial approximation. Originally developed to improve upon the accuracy of Gaussian elimination, interest in IR has been revived because of its suitability for execution on fast low-precision hardware such as analog devices and graphics processing units. IR generally converges when the error associated with the solution method is small, but is known to diverge when this error is large. We propose and analyze a novel enhancement to the IR algorithm by adding a line search optimization step that guarantees the algorithm will not diverge. Numerical experiments verify our theoretical results and illustrate the effectiveness of our proposed scheme.
format Preprint
id arxiv_https___arxiv_org_abs_2309_07865
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Stable iterative refinement algorithms for solving linear systems
Wu, Chai Wah
Squillante, Mark S.
Kalantzis, Vasileios
Horesh, Lior
Numerical Analysis
65F10
G.1.3
Iterative refinement (IR) is a popular scheme for solving a linear system of equations based on gradually improving the accuracy of an initial approximation. Originally developed to improve upon the accuracy of Gaussian elimination, interest in IR has been revived because of its suitability for execution on fast low-precision hardware such as analog devices and graphics processing units. IR generally converges when the error associated with the solution method is small, but is known to diverge when this error is large. We propose and analyze a novel enhancement to the IR algorithm by adding a line search optimization step that guarantees the algorithm will not diverge. Numerical experiments verify our theoretical results and illustrate the effectiveness of our proposed scheme.
title Stable iterative refinement algorithms for solving linear systems
topic Numerical Analysis
65F10
G.1.3
url https://arxiv.org/abs/2309.07865