Predicting performance of SMP clusters

kapitola v knize

angličtina

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.

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

ČEJKA, R.; DVOŘÁK, V.; STAROBA, J.

2002

2. 9. 2002

Kluwer Academic Publishers

Boston/London

1-4020-7209-0

Distributed and Parallel Systems - Cluster and Grid Computing

38

45

8