Buscar

 

RI UFLA (Universidade Federal de Lavras) >
Revistas UFLA >
Infocomp >

Please use this identifier to cite or link to this item: http://repositorio.ufla.br/jspui/handle/1/10033

Title: Performance analysis of static constraint network partitioning algorithm using multithreads
???metadata.dc.creator???: Veiga, Rolf Pagotto
Pereira, Marluce Rodrigues
Lacerda, Wilian Soares
Keywords: Multithread programming
Constraint networks
Programação Multithread
Redes de restrição
Publisher: Editora da UFLA
???metadata.dc.date???: 1-Jul-2010
Citation: VEIGA, R. P.; PEREIRA, M. R.; LACERDA, W. S. Performance analysis of static constraint network partitioning algorithm using multithreads. INFOCOMP: Journal of Computer Science, Lavras, v. 9, n. 6, p. 34-42, July 2010. (Special Issue).
Abstract: The multicore processors can have its processing capacity well used when the applications are implemented in parallel using multithread programming. So, problems which demand high performance capacity and high data quantity can be executed faster. This work presents a performance analyze of a static partitioning algorithm over applications modeled in constraint graphs, named Grouping-Sink [2], in different hardware and software platforms. The sequential algorithm was parallelized using multithreads. The experiments were done over constraint networks. The results show that the parallelization using multithreads can achieve speedups until 473.19, using Linux operating system, with high priority. Besides, the threads scheduling realized by the operating system and the hardware architecture characteristics can affect the execution time of a parallel application.
Other Identifiers: http://www.dcc.ufla.br/infocomp/index.php/INFOCOMP/article/view/382
???metadata.dc.language???: eng
Appears in Collections:Infocomp

Files in This Item:

There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


View Statistics

 


DSpace Software Copyright © 2002-2010  Duraspace - Feedback