A robust and efficient parallel SVD solver based on restarted Lanczos bidiagonalization
Electronic transactions on numerical analysis, Tome 31 (2008), pp. 68-85.

Voir la notice de l'article provenant de la source Electronic Library of Mathematics

Summary: Lanczos bidiagonalization is a competitive method for computing a partial singular value decomposition of a large sparse matrix, that is, when only a subset of the singular values and corresponding singular vectors are required. However, a straightforward implementation of the algorithm has the problem of loss of orthogonality between computed Lanczos vectors, and some reorthogonalization technique must be applied. Also, an effective restarting strategy must be used to prevent excessive growth of the cost of reorthogonalization per iteration. On the other hand, if the method is to be implemented on a distributed-memory parallel computer, then additional precautions are required so that parallel efficiency is maintained as the number of processors increases.
Classification : 65F15, 15A18, 65F50
Keywords: partial singular value decomposition, Lanczos bidiagonalization, thick restart, parallel computing
@article{ETNA_2008__31__a17,
     author = {Hern\'andez, Vicente and Rom\'an, Jos\'e E. and Tom\'as, Andr\'es},
     title = {A robust and efficient parallel {SVD} solver based on restarted {Lanczos} bidiagonalization},
     journal = {Electronic transactions on numerical analysis},
     pages = {68--85},
     publisher = {mathdoc},
     volume = {31},
     year = {2008},
     language = {en},
     url = {http://geodesic.mathdoc.fr/item/ETNA_2008__31__a17/}
}
TY  - JOUR
AU  - Hernández, Vicente
AU  - Román, José E.
AU  - Tomás, Andrés
TI  - A robust and efficient parallel SVD solver based on restarted Lanczos bidiagonalization
JO  - Electronic transactions on numerical analysis
PY  - 2008
SP  - 68
EP  - 85
VL  - 31
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/ETNA_2008__31__a17/
LA  - en
ID  - ETNA_2008__31__a17
ER  - 
%0 Journal Article
%A Hernández, Vicente
%A Román, José E.
%A Tomás, Andrés
%T A robust and efficient parallel SVD solver based on restarted Lanczos bidiagonalization
%J Electronic transactions on numerical analysis
%D 2008
%P 68-85
%V 31
%I mathdoc
%U http://geodesic.mathdoc.fr/item/ETNA_2008__31__a17/
%G en
%F ETNA_2008__31__a17
Hernández, Vicente; Román, José E.; Tomás, Andrés. A robust and efficient parallel SVD solver based on restarted Lanczos bidiagonalization. Electronic transactions on numerical analysis, Tome 31 (2008), pp. 68-85. http://geodesic.mathdoc.fr/item/ETNA_2008__31__a17/