Empirical parallel performance prediction from Semantics-based profiling

Norman Scaife; Greg Michaelson; Susumu Horiguchi

doi:10.1007/11428848_100

Empirical parallel performance prediction from Semantics-based profiling

Norman Scaife, Greg Michaelson, Susumu Horiguchi

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

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 language	English
Title of host publication	Computational Science – ICCS 2005
Subtitle of host publication	5th International Conference, Atlanta, GA, USA, May 22-25, 2005. Proceedings, Part II
Pages	781-789
Number of pages	9
Volume	3515
ISBN (Electronic)	978-3-540-32114-9
DOIs	https://doi.org/10.1007/11428848_100
Publication status	Published - 2005
Event	5th International Conference on Computational Science - Atlanta, GA, United States Duration: 22 May 2005 → 25 May 2005

Publication series

Name	Lecture Notes in Computer Science
Volume	3515
ISSN (Print)	0302-9743

Conference

Conference	5th International Conference on Computational Science
Abbreviated title	ICCS 2005
Country/Territory	United States
City	Atlanta, GA
Period	22/05/05 → 25/05/05

Access to Document

10.1007/11428848_100

Cite this

@inbook{3574ce02b7384c11a1c1350318a65f88,

title = "Empirical parallel performance prediction from Semantics-based profiling",

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. {\textcopyright} Springer-Verlag Berlin Heidelberg 2005.",

author = "Norman Scaife and Greg Michaelson and Susumu Horiguchi",

year = "2005",

doi = "10.1007/11428848\_100",

language = "English",

isbn = "978-3-540-26043-1",

volume = "3515",

series = "Lecture Notes in Computer Science",

pages = "781--789",

booktitle = "Computational Science – ICCS 2005",

note = "5th International Conference on Computational Science, ICCS 2005 ; Conference date: 22-05-2005 Through 25-05-2005",

}

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, Lecture Notes in Computer Science, vol. 3515, pp. 781-789, 5th International Conference on Computational Science, Atlanta, GA, United States, 22/05/05. https://doi.org/10.1007/11428848_100

Empirical parallel performance prediction from Semantics-based profiling. / Scaife, Norman; Michaelson, Greg; Horiguchi, Susumu.
Computational Science – ICCS 2005: 5th International Conference, Atlanta, GA, USA, May 22-25, 2005. Proceedings, Part II. Vol. 3515 2005. p. 781-789 (Lecture Notes in Computer Science; Vol. 3515).

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - Empirical parallel performance prediction from Semantics-based profiling

AU - Scaife, Norman

AU - Michaelson, Greg

AU - Horiguchi, Susumu

PY - 2005

Y1 - 2005

N2 - 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.

AB - 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.

UR - https://www.scopus.com/pages/publications/25144470170

U2 - 10.1007/11428848_100

DO - 10.1007/11428848_100

M3 - Chapter (peer-reviewed)

SN - 978-3-540-26043-1

VL - 3515

T3 - Lecture Notes in Computer Science

SP - 781

EP - 789

BT - Computational Science – ICCS 2005

T2 - 5th International Conference on Computational Science

Y2 - 22 May 2005 through 25 May 2005

ER -

Empirical parallel performance prediction from Semantics-based profiling

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this