** Next:** Lanczos Method A.
** Up:** Single- and Multiple-Vector Iterations
** Previous:** Acceleration.
** Contents**
** Index**

There are several pieces of software available for subspace
iteration. EA12 is a routine by Duff and Scott for symmetric
subspace iteration [143]. It is part of the Harwell
Subroutine Library. Given a real (sparse) symmetric matrix,
this routine computes the largest eigenvalues and their
corresponding eigenvectors. There is a classical
implementation of subspace iteration by Rutishauser [382]
in [458], where most of the refinements discussed
here have been included.
There are also two subspace iteration-based software
packages for the nonsymmetric eigenvalue problem, SRRIT by
Bai and Stewart [37] and LOPSI by Stewart and Jennings [427]
(see §7.4 for more details).
These packages can be used to
solve the symmetric eigenvalue problem, but with some loss
in efficiency and accuracy.

For more information about this software, including how to access
it, see the book's homepage ETHOME.

Susan Blackford
2000-11-20