Block Tridiagonal Solvers on Heterogeneous Architectures

Author

Valero-Lara, Pedro ; Pinelli, Alfredo ; Favier, Julien ; Matias, Manuel Prieto

Author_Institution

Unidad de Modelizacion y Simulacion Numerica, CIEMAT, Madrid, Spain

fYear

2012

fDate

10-13 July 2012

Firstpage

609

Lastpage

616

Abstract

Modern multi-core and many-core systems offer a very impressive cost/performance ratio. In this paper a set of new parallel implementations for the solution of linear systems with block-tridiagonal coefficient matrix on current parallel architectures is proposed and evaluated: one of them on multi-core, others on many-core and finally, a new heterogeneous implementation on both architectures. The results show a speedup higher than 6 on certain parts of the problem, being the heterogeneous implementation the fastest.

Keywords

matrix algebra; multiprocessing systems; parallel architectures; block tridiagonal coefficient matrix; block tridiagonal solvers; heterogeneous architectures; many core systems; multicore systems; parallel architectures; parallel implementations; Computer architecture; Equations; Graphics processing unit; Instruction sets; Linear systems; Parallel processing; Vectors; Block-tridiagonal linear systems; multi-core and many-core; parallel processing;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Processing with Applications (ISPA), 2012 IEEE 10th International Symposium on

Conference_Location

Leganes

Print_ISBN

978-1-4673-1631-6

Type

conf

DOI

10.1109/ISPA.2012.91

Filename

6280351