Publication Details

Predicting performance of SMP clusters

ČEJKA Rudolf, DVOŘÁK Václav and STAROBA Jiří. Predicting performance of SMP clusters. Distributed and Parallel Systems - Cluster and Grid Computing. Boston/London: Kluwer Academic Publishers, 2002, pp. 38-45. ISBN 1-4020-7209-0.
Czech title
Predikce výkonnosti svazků symetrických multiprocesorů
Type
book chapter
Language
english
Authors
Čejka Rudolf, Ing. (CC FIT BUT)
Dvořák Václav, prof. Ing., DrSc. (DCSY FIT BUT)
Staroba Jiří, Ing. (DCSY FIT BUT)
Keywords

SMP clusters, performance modeling, performance prediction, communicating processes, hybrid parallel programming model

Abstract

The paper reports on simulation-based performance prediction of parallel applications running on SMP clusters. First the typical message-passing programs are investigated, which demonstrate different computation/ communication patterns: process farm, pipeline and data parallel computing on a linear processor chain. The usefulness of various SMP cluster configurations is discussed with regard to obtained results. Then the message-passing model of shared memory and SMP clusters is used to simulate execution of hybrid message-passing/shared objects programs. A large system of linear equations is used as an example. A single SMP, a SMP cluster and a workstation cluster are compared in performance with a limited L1-cache size.

Published
2002
Pages
38-45
Book
Distributed and Parallel Systems - Cluster and Grid Computing
ISBN
1-4020-7209-0
Publisher
Kluwer Academic Publishers
Place
Boston/London, US
BibTeX 
@INBOOK{FITPUB6950,
   author = "Rudolf \v{C}ejka and V\'{a}clav Dvo\v{r}\'{a}k and Ji\v{r}\'{i} Staroba",
   title = "Predicting performance of SMP clusters",
   pages = "38--45",
   booktitle = " Distributed and Parallel Systems - Cluster and Grid Computing",
   year = 2002,
   location = "Boston/London, US",
   publisher = "Kluwer Academic Publishers",
   ISBN = "1-4020-7209-0",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/6950"
}
Back to top