文档介绍：Journal putational and Applied Mathematics 123 (2000) 489–514
ate/cam
Numerical linear algebra algorithms and software
Jack J. Dongarraa;b;∗, Victor Eijkhouta
aDepartment puter Science, University of Tennessee, 107 Ayres Hall, Knoxville, TN 37996-1301, USA
bMathematical Sciences Section, Oak Ridge National Laboratory, . Box 2008, Bldg. 6012, Oak Ridge,
TN 37831-6367, USA
Received 12 July 1999; received in revised form 16 August 1999
Abstract
The increasing availability of advanced-puters has a signiÿcant e ect on all spheres of pu-
tation, including algorithm research and software development in numerical linear algebra. Linear algebra – in particular,
the solution of linear systems of equations – lies at the heart of most calculations in puting. This paper
discusses some of the recent developments in linear algebra designed to exploit these advanced-puters.
We discuss two broad classes of algorithms: those for dense, and those for sparse matrices.
c 2000 Elsevier Science
. All rights reserved.
1. Introduction
The increasing availability of advanced-puters has a signiÿcant e ect on all spheres
of putation, including algorithm research and software development in numerical linear
algebra. Linear algebra – in particular, the solution of linear systems of equations – lies at the heart
of most calculations in puting. This article discusses some of the recent developments
in linear algebra designed to exploit these advanced-puters. We discuss two broad
classes of algorithms: those for dense, and those for sparse matrices. A matrix is called sparse if it
has a substantial number of zero elements, making specialized storage and algorithms necessary.
Much of the work in developing linear algebra software for advanced-puters is
motivated by the need to solve large problems on the puters available. In this article,
we focus on four basic issues: (1) the motivation for the work; (2) the development of standards
∗ Corresponding author