Methodology for the performance prediction of massively parallel applications

Daniel Menasce; Sam H. Noh; Satish K. Tripathi

Methodology for the performance prediction of massively parallel applications

Daniel Menasce
, Sam H. Noh
, Satish K. Tripathi

OTP Administrative Operations

University at Buffalo

George Mason University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

This paper presents a methodology to predict the execution time of massively parallel applications before any significant implementation actions are taken. This methodology captures the problem decomposition into tasks and their precedence relationship, along with the computational and communication demands placed by the application on the underlying architecture. An example shows how the methodology may be used to study the effects of various data placement strategies, problem size, and number of processors for an LU factorization algorithm. The model predictions were validated with published experimental results on a Touchstone Delta machine.

Original language	English
Title of host publication	Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing
Editors	Anon
Publisher	Publ by IEEE
Pages	250-257
Number of pages	8
ISBN (Print)	081864222X
State	Published - 1993
Event	Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing - Dallas, TX, USA Duration: Dec 1 1993 → Dec 4 1993

Publication series

Name	Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing

Conference

Conference	Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing
City	Dallas, TX, USA
Period	12/1/93 → 12/4/93

Cite this

@inproceedings{2d85ee8b0c2842f1a536614254f29153,

title = "Methodology for the performance prediction of massively parallel applications",

abstract = "This paper presents a methodology to predict the execution time of massively parallel applications before any significant implementation actions are taken. This methodology captures the problem decomposition into tasks and their precedence relationship, along with the computational and communication demands placed by the application on the underlying architecture. An example shows how the methodology may be used to study the effects of various data placement strategies, problem size, and number of processors for an LU factorization algorithm. The model predictions were validated with published experimental results on a Touchstone Delta machine.",

author = "Daniel Menasce and Noh, \{Sam H.\} and Tripathi, \{Satish K.\}",

year = "1993",

language = "English",

isbn = "081864222X",

series = "Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing",

publisher = "Publ by IEEE",

pages = "250--257",

editor = "Anon",

booktitle = "Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing",

note = "Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing ; Conference date: 01-12-1993 Through 04-12-1993",

}

Menasce, D, Noh, SH & Tripathi, SK 1993, Methodology for the performance prediction of massively parallel applications. in Anon (ed.), Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing. Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing, Publ by IEEE, pp. 250-257, Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing, Dallas, TX, USA, 12/1/93.

Methodology for the performance prediction of massively parallel applications. / Menasce, Daniel; Noh, Sam H.; Tripathi, Satish K.
Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing. ed. / Anon. Publ by IEEE, 1993. p. 250-257 (Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Methodology for the performance prediction of massively parallel applications

AU - Menasce, Daniel

AU - Noh, Sam H.

AU - Tripathi, Satish K.

PY - 1993

Y1 - 1993

N2 - This paper presents a methodology to predict the execution time of massively parallel applications before any significant implementation actions are taken. This methodology captures the problem decomposition into tasks and their precedence relationship, along with the computational and communication demands placed by the application on the underlying architecture. An example shows how the methodology may be used to study the effects of various data placement strategies, problem size, and number of processors for an LU factorization algorithm. The model predictions were validated with published experimental results on a Touchstone Delta machine.

AB - This paper presents a methodology to predict the execution time of massively parallel applications before any significant implementation actions are taken. This methodology captures the problem decomposition into tasks and their precedence relationship, along with the computational and communication demands placed by the application on the underlying architecture. An example shows how the methodology may be used to study the effects of various data placement strategies, problem size, and number of processors for an LU factorization algorithm. The model predictions were validated with published experimental results on a Touchstone Delta machine.

UR - https://www.scopus.com/pages/publications/0027810687

M3 - Conference contribution

SN - 081864222X

T3 - Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing

SP - 250

EP - 257

BT - Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing

A2 - Anon, null

PB - Publ by IEEE

T2 - Proceedings of the 5th IEEE Symposium on Parallel and Distributed Processing

Y2 - 1 December 1993 through 4 December 1993

ER -

Methodology for the performance prediction of massively parallel applications

Abstract

Publication series

Conference

Other files and links

Fingerprint

Cite this