User: Guest  Login
Title:

Parallel Eigenvalue Computation for Banded Generalized Eigenvalue Problems

Document type:
Zeitschriftenaufsatz
Author(s):
Rippl, Michael; Lang, Bruno; Huckle, Thomas
Abstract:
We consider generalized eigenvalue problems $Ax=Bx\lambda$ with banded hermitian matrix $A$ and hermitian positive definite $B$. To reduce the generalized eigenvalue problem to standard form $Cy=y\lambda$ the algorithm proposed by Crawford is applied preserving the banded structure in $C$. We present a parallel implementation of this method included in the ELPA library. Numerical experiments show the advantages of this approach compared to standard solvers.
Congress title:
International Workshop on Parallel Matrix Algorithms and Applications
Journal title:
Parallel Computing
Year:
2019
Journal volume:
88
Year / month:
2019-04
Quarter:
2. Quartal
Month:
Apr
Pages contribution:
102542
Reviewed:
ja
Language:
en
Fulltext / DOI:
doi:10.1016/j.parco.2019.07.002
WWW:
https://www.sciencedirect.com/science/article/pii/S0167819119301279
Publisher:
Elsevier
Status:
Verlagsversion / published
Accepted:
29.07.2019
Date of publication:
04.08.2019
TUM Institution:
Department of Informatics
 BibTeX