Lanczos Algorithms for Large Symmetric Eigenvalue Computations. Volume 2. Documentation and Listings. Original Lanczos Codes

Издательство Birkhäuser, 1985, -600 pp.Volume 2 of this book (/file/725043/), which is not being reproduced in the SIAM Classics in Applied Mathematics series, contains listings of all of the original implementations of the algorithms that are discussed in this monograph. The text for Volume 2, along with the source code for each of those listings, is available at the numerical analysis community Web repository, www.netlib.org, under the name "lanczos."
We hope that republication of this book will be of benefit to the computational physics, chemistry, and engineering communities that have shown interest in these types of algorithms. We also hope that students in the numerical linear algebra community will find this book useful as a means for building appreciation for the potential gains that may be achievable by thinking outside of the orthogonal box.
This book was published 17 years ago. Computers of today are orders of magnitude faster and have orders of magnitude more memory and storage than those which were available when this book was written. Seventeen years ago, a matrix of size 10, 000 was considered very large. Since 1985 some of the algorithms which are included in this book have been used on problems of size a million or more. The requirements are accurate matrix computations and sufficient computer arithmetic precision.Lanczos procedures
Real Symmetric Matrices
Hermitian Matrices
Factored Inverses of Real Symmetric Matrices
Real Symmetric Generalized Problems
Real Rectangular Matrices, Singular Values and Vectors
Nondefective Complex Symmetric Matrices
Real Symmetric Matrices, Block Lanczos Code
Factored Inverses, Real Symmetric Block Lanczos Code
Errata: Volume I: TheoryVolume 2 of the book (/file/725043/), which is not being reproduced in the SIAM Classics in Applied Mathematics series, contains listings of all of the original implementations of the algorithms that are discussed in this monograph. The text for Volume 2, along with the source code for each of those listings, is available at the numerical analysis community Web repository, www.netlib.org, under the name.