Cargando…

Comparison of eigensolvers for symmetric band matrices

We compare different algorithms for computing eigenvalues and eigenvectors of a symmetric band matrix across a wide range of synthetic test problems. Of particular interest is a comparison of state-of-the-art tridiagonalization-based methods as implemented in Lapack or Plasma on the one hand, and th...

Descripción completa

Detalles Bibliográficos
Autores principales: Moldaschl, Michael, Gansterer, Wilfried N.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Elsevier B.V 2014
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4617464/
https://www.ncbi.nlm.nih.gov/pubmed/26594079
http://dx.doi.org/10.1016/j.scico.2014.01.005
_version_ 1782396799478136832
author Moldaschl, Michael
Gansterer, Wilfried N.
author_facet Moldaschl, Michael
Gansterer, Wilfried N.
author_sort Moldaschl, Michael
collection PubMed
description We compare different algorithms for computing eigenvalues and eigenvectors of a symmetric band matrix across a wide range of synthetic test problems. Of particular interest is a comparison of state-of-the-art tridiagonalization-based methods as implemented in Lapack or Plasma on the one hand, and the block divide-and-conquer (BD&C) algorithm as well as the block twisted factorization (BTF) method on the other hand. The BD&C algorithm does not require tridiagonalization of the original band matrix at all, and the current version of the BTF method tridiagonalizes the original band matrix only for computing the eigenvalues. Avoiding the tridiagonalization process sidesteps the cost of backtransformation of the eigenvectors. Beyond that, we discovered another disadvantage of the backtransformation process for band matrices: In several scenarios, a lot of gradual underflow is observed in the (optional) accumulation of the transformation matrix and in the (obligatory) backtransformation step. According to the IEEE 754 standard for floating-point arithmetic, this implies many operations with subnormal (denormalized) numbers, which causes severe slowdowns compared to the other algorithms without backtransformation of the eigenvectors. We illustrate that in these cases the performance of existing methods from Lapack and Plasma reaches a competitive level only if subnormal numbers are disabled (and thus the IEEE standard is violated). Overall, our performance studies illustrate that if the problem size is large enough relative to the bandwidth, BD&C tends to achieve the highest performance of all methods if the spectrum to be computed is clustered. For test problems with well separated eigenvalues, the BTF method tends to become the fastest algorithm with growing problem size.
format Online
Article
Text
id pubmed-4617464
institution National Center for Biotechnology Information
language English
publishDate 2014
publisher Elsevier B.V
record_format MEDLINE/PubMed
spelling pubmed-46174642015-11-20 Comparison of eigensolvers for symmetric band matrices Moldaschl, Michael Gansterer, Wilfried N. Sci Comput Program Article We compare different algorithms for computing eigenvalues and eigenvectors of a symmetric band matrix across a wide range of synthetic test problems. Of particular interest is a comparison of state-of-the-art tridiagonalization-based methods as implemented in Lapack or Plasma on the one hand, and the block divide-and-conquer (BD&C) algorithm as well as the block twisted factorization (BTF) method on the other hand. The BD&C algorithm does not require tridiagonalization of the original band matrix at all, and the current version of the BTF method tridiagonalizes the original band matrix only for computing the eigenvalues. Avoiding the tridiagonalization process sidesteps the cost of backtransformation of the eigenvectors. Beyond that, we discovered another disadvantage of the backtransformation process for band matrices: In several scenarios, a lot of gradual underflow is observed in the (optional) accumulation of the transformation matrix and in the (obligatory) backtransformation step. According to the IEEE 754 standard for floating-point arithmetic, this implies many operations with subnormal (denormalized) numbers, which causes severe slowdowns compared to the other algorithms without backtransformation of the eigenvectors. We illustrate that in these cases the performance of existing methods from Lapack and Plasma reaches a competitive level only if subnormal numbers are disabled (and thus the IEEE standard is violated). Overall, our performance studies illustrate that if the problem size is large enough relative to the bandwidth, BD&C tends to achieve the highest performance of all methods if the spectrum to be computed is clustered. For test problems with well separated eigenvalues, the BTF method tends to become the fastest algorithm with growing problem size. Elsevier B.V 2014-09-15 /pmc/articles/PMC4617464/ /pubmed/26594079 http://dx.doi.org/10.1016/j.scico.2014.01.005 Text en © 2014 The Authors http://creativecommons.org/licenses/by-nc-nd/3.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
spellingShingle Article
Moldaschl, Michael
Gansterer, Wilfried N.
Comparison of eigensolvers for symmetric band matrices
title Comparison of eigensolvers for symmetric band matrices
title_full Comparison of eigensolvers for symmetric band matrices
title_fullStr Comparison of eigensolvers for symmetric band matrices
title_full_unstemmed Comparison of eigensolvers for symmetric band matrices
title_short Comparison of eigensolvers for symmetric band matrices
title_sort comparison of eigensolvers for symmetric band matrices
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4617464/
https://www.ncbi.nlm.nih.gov/pubmed/26594079
http://dx.doi.org/10.1016/j.scico.2014.01.005
work_keys_str_mv AT moldaschlmichael comparisonofeigensolversforsymmetricbandmatrices
AT ganstererwilfriedn comparisonofeigensolversforsymmetricbandmatrices