Predictive performance modelling of parallel component compositions

Lei Zhao; Stephen A. Jarvis

doi:10.1007/s10586-007-0010-2

Predictive performance modelling of parallel component compositions

Lei Zhao^*, Stephen A. Jarvis

^*Corresponding author for this work

Engineering and Physical Sciences

Research output: Contribution to journal › Article › peer-review

Abstract

Large-scale scientific computing applications frequently make use of closely-coupled distributed parallel components. The performance of such applications is therefore dependent on the component parts and their interaction at run-time. This paper describes a methodology for predictive performance modelling and evaluation of parallel applications composed of multiple interacting components. In this paper, the fundamental steps and required operations involved in the modelling and evaluation process are identified - including component decomposition, component model combination, M × N communication modelling, dataflow analysis and overall performance evaluation. A case study is presented to illustrate the modelling process and the methodology is verified through experimental analysis.

Original language	English
Pages (from-to)	155-166
Number of pages	12
Journal	Cluster Computing
Volume	10
Issue number	2
Early online date	24 Mar 2007
DOIs	https://doi.org/10.1007/s10586-007-0010-2
Publication status	Published - Jun 2007

Keywords

M × N communication
Parallel component composition
Performance modelling

ASJC Scopus subject areas

Software
Computer Networks and Communications

Access to Document

10.1007/s10586-007-0010-2Licence: None: All rights reserved

Cite this

@article{71a36966bbe041cbad1b85f6d888421b,

title = "Predictive performance modelling of parallel component compositions",

abstract = "Large-scale scientific computing applications frequently make use of closely-coupled distributed parallel components. The performance of such applications is therefore dependent on the component parts and their interaction at run-time. This paper describes a methodology for predictive performance modelling and evaluation of parallel applications composed of multiple interacting components. In this paper, the fundamental steps and required operations involved in the modelling and evaluation process are identified - including component decomposition, component model combination, M × N communication modelling, dataflow analysis and overall performance evaluation. A case study is presented to illustrate the modelling process and the methodology is verified through experimental analysis.",

keywords = "M × N communication, Parallel component composition, Performance modelling",

author = "Lei Zhao and Jarvis, {Stephen A.}",

year = "2007",

month = jun,

doi = "10.1007/s10586-007-0010-2",

language = "English",

volume = "10",

pages = "155--166",

journal = "Cluster Computing",

issn = "1386-7857",

publisher = "Springer",

number = "2",

}

TY - JOUR

T1 - Predictive performance modelling of parallel component compositions

AU - Zhao, Lei

AU - Jarvis, Stephen A.

PY - 2007/6

Y1 - 2007/6

N2 - Large-scale scientific computing applications frequently make use of closely-coupled distributed parallel components. The performance of such applications is therefore dependent on the component parts and their interaction at run-time. This paper describes a methodology for predictive performance modelling and evaluation of parallel applications composed of multiple interacting components. In this paper, the fundamental steps and required operations involved in the modelling and evaluation process are identified - including component decomposition, component model combination, M × N communication modelling, dataflow analysis and overall performance evaluation. A case study is presented to illustrate the modelling process and the methodology is verified through experimental analysis.

AB - Large-scale scientific computing applications frequently make use of closely-coupled distributed parallel components. The performance of such applications is therefore dependent on the component parts and their interaction at run-time. This paper describes a methodology for predictive performance modelling and evaluation of parallel applications composed of multiple interacting components. In this paper, the fundamental steps and required operations involved in the modelling and evaluation process are identified - including component decomposition, component model combination, M × N communication modelling, dataflow analysis and overall performance evaluation. A case study is presented to illustrate the modelling process and the methodology is verified through experimental analysis.

KW - M × N communication

KW - Parallel component composition

KW - Performance modelling

UR - http://www.scopus.com/inward/record.url?scp=34248656746&partnerID=8YFLogxK

U2 - 10.1007/s10586-007-0010-2

DO - 10.1007/s10586-007-0010-2

M3 - Article

AN - SCOPUS:34248656746

SN - 1386-7857

VL - 10

SP - 155

EP - 166

JO - Cluster Computing

JF - Cluster Computing

IS - 2

ER -

Predictive performance modelling of parallel component compositions

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this