Evaluating the hyperbolic model on a variety of architectures

Ion Stoica, Florin Sultan, David Keyes

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

We illustrate the application of the hyperbolic model, which generalizes standard two-parameter dedicated-link models for communication costs in message-passing environments, to four distributed-memory architectures: Ethernet NOW, FDDI NOW, IBM SP2, and Intel Paragon. We first evaluate the parameters of the model from simple communication patterns. Then overall communication time estimates, which compare favorably with experimental measurements, are deduced for the message traffic in a scientific application code. For transformational computing on dedicated systems, for which message traffic is describable in terms of a finite number of regular patterns, the model offers a good compromise between the competing objectives of flexibility, tractability, and reliability of prediction.

Original languageEnglish (US)
Title of host publicationEuro-Par 1996 Parallel Processing - 2nd International Euro-Par Conference, Proceedings
EditorsLuc Bouge, Pierre Fraigniaud, Anne Mignotte, Yves Robert
PublisherSpringer Verlag
Pages387-396
Number of pages10
ISBN (Print)3540616276, 9783540616276
DOIs
StatePublished - Jan 1 1996
Event2nd International Euro-Par Conference on Parallel Processing, Euro-Par 1996 - Lyon, France
Duration: Aug 26 1996Aug 29 1996

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1124
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2nd International Euro-Par Conference on Parallel Processing, Euro-Par 1996
CountryFrance
CityLyon
Period08/26/9608/29/96

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Evaluating the hyperbolic model on a variety of architectures'. Together they form a unique fingerprint.

Cite this