Tall-and-skinny QR factorization with approximate Householder reflectors on graphics processors
Impacto
Scholar |
Otros documentos de la autoría: Tomás Domínguez, Andrés Enrique; Quintana-Orti, Enrique S.
Metadatos
Mostrar el registro completo del ítemcomunitat-uji-handle:10234/9
comunitat-uji-handle2:10234/7036
comunitat-uji-handle3:10234/8620
comunitat-uji-handle4:
INVESTIGACIONEste recurso está restringido
https://doi.org/10.1007/s11227-020-03176-3 |
Metadatos
Título
Tall-and-skinny QR factorization with approximate Householder reflectors on graphics processorsFecha de publicación
2020-01-24Editor
SpringerCita bibliográfica
Tomás, A.E., Quintana-Ortí, E.S. Tall-and-skinny QR factorization with approximate Householder reflectors on graphics processors. J Supercomput (2020). https://doi.org/10.1007/s11227-020-03176-3Tipo de documento
info:eu-repo/semantics/articleVersión de la editorial
https://link.springer.com/article/10.1007%2Fs11227-020-03176-3Versión
info:eu-repo/semantics/publishedVersionPalabras clave / Materias
Resumen
We present a novel method for the QR factorization of large tall-and-skinny matrices that introduces an approximation technique for computing the Householder vectors. This approach is very competitive on a hybrid ... [+]
We present a novel method for the QR factorization of large tall-and-skinny matrices that introduces an approximation technique for computing the Householder vectors. This approach is very competitive on a hybrid platform equipped with a graphics processor, with a performance advantage over the conventional factorization due to the reduced amount of data transfers between the graphics accelerator and the main memory of the host. Our experiments show that, for tall–skinny matrices, the new approach outperforms the code in MAGMA by a large margin, while it is very competitive for square matrices when the memory transfers and CPU computations are the bottleneck of the Householder QR factorization. [-]
Proyecto de investigación
Project TIN2017-82972-R from the MINECO (Spain) ; EU H2020 Project 732631 “OPRECOMP. Open Transprecision Computing”.Derechos de acceso
© Springer Science+Business Media, LLC, part of Springer Nature 2020
http://rightsstatements.org/vocab/InC/1.0/
info:eu-repo/semantics/restrictedAccess
http://rightsstatements.org/vocab/InC/1.0/
info:eu-repo/semantics/restrictedAccess
Aparece en las colecciones
- ICC_Articles [419]