Lanczos methods for the smallest eigenvalues of large matrices on distributed memory supercomputers
Conference
·
OSTI ID:54447
- Western Michigan Univ., Kalamazoo, MI (United States)
We present results on applying Lanczos methods to find some of the eigenvalues of dense matrices whose size renders reduction to fill tridiagonal or hessenberg form undesirable. The use of distributed memory supercomputers is an ideal match for many problem in physics and other applications, where a system is modeled by building a large matrix, whose entries must be computed and whose few smallest eigenvalues provide the desired information. Our motivating application starts with 2000 by 2000 systems.
- OSTI ID:
- 54447
- Report Number(s):
- DOE/ER/25151--1-Vol.1; CONF-930331--Vol.1
- Country of Publication:
- United States
- Language:
- English
Similar Records
Lanczos algorithm for symmetric eigenvalue problems
The BR eigenvalue algorithm
Finding eigenvalues and eigenvectors of unsymmetric matrices using a distributed-memory multiprocessor
Conference
·
Mon Dec 31 23:00:00 EST 1979
·
OSTI ID:5171549
The BR eigenvalue algorithm
Conference
·
Fri Oct 31 23:00:00 EST 1997
·
OSTI ID:672016
Finding eigenvalues and eigenvectors of unsymmetric matrices using a distributed-memory multiprocessor
Technical Report
·
Mon Oct 31 23:00:00 EST 1988
·
OSTI ID:6637340