Empirical parallel performance prediction from Semantics-based profiling

Norman Scaife, Greg Michaelson, Susumu Horiguchi

Research output: Chapter in Book/Report/Conference proceedingChapter (peer-reviewed)

1 Citation (Scopus)

Abstract

The PMLS parallelizing compiler for Standard ML is based upon the automatic instantiation of algorithmic skeletons at sites of higher order function use. PMLS seeks to optimise run-time parallel behaviour by combining skeleton cost models with Structural Operational Semantics rule counts for HOF argument functions. In this paper, the formulation of a general rule count cost model as a set of over-determined linear equations is discussed, and their solution by singular value decomposition, and by a genetic algorithm, are presented. © Springer-Verlag Berlin Heidelberg 2005.

Original languageEnglish
Title of host publicationComputational Science – ICCS 2005
Subtitle of host publication5th International Conference, Atlanta, GA, USA, May 22-25, 2005. Proceedings, Part II
Pages781-789
Number of pages9
Volume3515
ISBN (Electronic)978-3-540-32114-9
DOIs
Publication statusPublished - 2005
Event5th International Conference on Computational Science - Atlanta, GA, United States
Duration: 22 May 200525 May 2005

Publication series

NameLecture Notes in Computer Science
Volume3515
ISSN (Print)0302-9743

Conference

Conference5th International Conference on Computational Science
Abbreviated titleICCS 2005
CountryUnited States
CityAtlanta, GA
Period22/05/0525/05/05

Fingerprint Dive into the research topics of 'Empirical parallel performance prediction from Semantics-based profiling'. Together they form a unique fingerprint.

  • Cite this

    Scaife, N., Michaelson, G., & Horiguchi, S. (2005). Empirical parallel performance prediction from Semantics-based profiling. In Computational Science – ICCS 2005: 5th International Conference, Atlanta, GA, USA, May 22-25, 2005. Proceedings, Part II (Vol. 3515, pp. 781-789). (Lecture Notes in Computer Science; Vol. 3515). https://doi.org/10.1007/11428848_100