On logic program parallélisation strategies

Research output: Chapter in Book/Report/Conference proceedingsChapterpeer-review

Abstract

Given an OR-paraUel logic program with several possible parallélisations, each with a good speedup, should we expect their combination to have a good speedup? Also, can a parallélisation with poor speedup ever have a good effect when combined with other parallélisations? This paper shows that combining speedups may have unexpected effects for both speculative and non-speculative computations, and that these effects have consequences for parallélisation strategies based on speedup measurements. The examples used for illustration axe also useful as pathological benchmarks for run-time and compile-time analysis tools, such as trace visualisers and automatic parallelisers.

Original languageEnglish
Title of host publicationParallel Processing
Subtitle of host publicationCONPAR 1994 - VAPP VI - 3rd Joint International Conference on Vector and Parallel Processing, Proceedings
EditorsBruno Buchberger, Jens Volkert
PublisherSpringer Verlag
Pages289-300
Number of pages12
ISBN (Print)9783540584308
DOIs
Publication statusPublished - 1994
Externally publishedYes
Event3rd Joint International Conference on Vector and Parallel Processing, CONPAR 1994 - VAPP VI - Linz, Austria
Duration: 6 Sep 19948 Sep 1994

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume854 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd Joint International Conference on Vector and Parallel Processing, CONPAR 1994 - VAPP VI
Country/TerritoryAustria
CityLinz
Period6/09/948/09/94

Fingerprint

Dive into the research topics of 'On logic program parallélisation strategies'. Together they form a unique fingerprint.

Cite this