Performance study and analysis of parallel multilevel preconditioners

Chi Shen; Jun Zhang

doi:10.1007/s10852-005-9026-x

Performance study and analysis of parallel multilevel preconditioners

Chi Shen
, Jun Zhang

Computer Science

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

Abstract

The significant gap between peak and realized performance of parallel systems motivates the need for performance analysis. In order to predict the performance of a class of parallel multilevel ILU preconditioner (PBILUM), we build two performance prediction models for both the preconditioner construction phase and the solution phase. These models combine theoretical features of the preconditioners with estimates on computation cost, communications overhead, etc. Experimental simulations show that our model predication based on certain reasonable assumptions is close to the simulation results. The models may be used to predict the performance of this class of parallel preconditioners.

Original language	English
Pages (from-to)	331-352
Number of pages	22
Journal	Journal of Mathematical Modelling and Algorithms
Volume	5
Issue number	3
DOIs	https://doi.org/10.1007/s10852-005-9026-x
State	Published - Sep 2006

Bibliographical note

Funding Information:
* The research work of the authors was supported in part by the U.S. National Science Foundation under grants CCR-9988165, CCR-0092532, ACR-0202934, and ACR-234270, by the U.S. Department of Energy Office of Science under grant DE-FG02-02ER45961, by the Kentucky Science & Engineering Foundation under grant KSEF-02-264-RED-002. j Corresponding author.

Funding

* The research work of the authors was supported in part by the U.S. National Science Foundation under grants CCR-9988165, CCR-0092532, ACR-0202934, and ACR-234270, by the U.S. Department of Energy Office of Science under grant DE-FG02-02ER45961, by the Kentucky Science & Engineering Foundation under grant KSEF-02-264-RED-002. j Corresponding author.

Funders	Funder number
U.S. Department of Energy Office of Basic Science	DE-FG02-02ER45961
U.S. National Science Foundation (NSF)	CCR-0092532, ACR-234270, ACR-0202934, CCR-9988165
Kentucky Science and Engineering Foundation	KSEF-02-264-RED-002

Keywords

Block independent set
Complexity models
Multilevel preconditioning
Parallel computing

ASJC Scopus subject areas

Modeling and Simulation
Applied Mathematics

Access to Document

10.1007/s10852-005-9026-x

Cite this

@article{03fb24b7030d4820887f00e9e673b801,

title = "Performance study and analysis of parallel multilevel preconditioners",

abstract = "The significant gap between peak and realized performance of parallel systems motivates the need for performance analysis. In order to predict the performance of a class of parallel multilevel ILU preconditioner (PBILUM), we build two performance prediction models for both the preconditioner construction phase and the solution phase. These models combine theoretical features of the preconditioners with estimates on computation cost, communications overhead, etc. Experimental simulations show that our model predication based on certain reasonable assumptions is close to the simulation results. The models may be used to predict the performance of this class of parallel preconditioners.",

keywords = "Block independent set, Complexity models, Multilevel preconditioning, Parallel computing",

author = "Chi Shen and Jun Zhang",

note = "Funding Information: * The research work of the authors was supported in part by the U.S. National Science Foundation under grants CCR-9988165, CCR-0092532, ACR-0202934, and ACR-234270, by the U.S. Department of Energy Office of Science under grant DE-FG02-02ER45961, by the Kentucky Science \& Engineering Foundation under grant KSEF-02-264-RED-002. j Corresponding author.",

year = "2006",

month = sep,

doi = "10.1007/s10852-005-9026-x",

language = "English",

volume = "5",

pages = "331--352",

number = "3",

}

TY - JOUR

T1 - Performance study and analysis of parallel multilevel preconditioners

AU - Shen, Chi

AU - Zhang, Jun

N1 - Funding Information: * The research work of the authors was supported in part by the U.S. National Science Foundation under grants CCR-9988165, CCR-0092532, ACR-0202934, and ACR-234270, by the U.S. Department of Energy Office of Science under grant DE-FG02-02ER45961, by the Kentucky Science & Engineering Foundation under grant KSEF-02-264-RED-002. j Corresponding author.

PY - 2006/9

Y1 - 2006/9

N2 - The significant gap between peak and realized performance of parallel systems motivates the need for performance analysis. In order to predict the performance of a class of parallel multilevel ILU preconditioner (PBILUM), we build two performance prediction models for both the preconditioner construction phase and the solution phase. These models combine theoretical features of the preconditioners with estimates on computation cost, communications overhead, etc. Experimental simulations show that our model predication based on certain reasonable assumptions is close to the simulation results. The models may be used to predict the performance of this class of parallel preconditioners.

AB - The significant gap between peak and realized performance of parallel systems motivates the need for performance analysis. In order to predict the performance of a class of parallel multilevel ILU preconditioner (PBILUM), we build two performance prediction models for both the preconditioner construction phase and the solution phase. These models combine theoretical features of the preconditioners with estimates on computation cost, communications overhead, etc. Experimental simulations show that our model predication based on certain reasonable assumptions is close to the simulation results. The models may be used to predict the performance of this class of parallel preconditioners.

KW - Block independent set

KW - Complexity models

KW - Multilevel preconditioning

KW - Parallel computing

UR - https://www.scopus.com/pages/publications/33744915757

UR - https://www.scopus.com/pages/publications/33744915757#tab=citedBy

U2 - 10.1007/s10852-005-9026-x

DO - 10.1007/s10852-005-9026-x

M3 - Article

AN - SCOPUS:33744915757

SN - 1570-1166

VL - 5

SP - 331

EP - 352

JO - Journal of Mathematical Modelling and Algorithms

JF - Journal of Mathematical Modelling and Algorithms

IS - 3

ER -

Performance study and analysis of parallel multilevel preconditioners

Abstract

Bibliographical note

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this