PARTITIONING STRATEGY FOR NONUNIFORM PROBLEMS ON MULTIPROCESSORS.

Marsha Berger, Shahid H. Bokhari

Research output: Contribution to journalArticle

Abstract

The authors consider the partitioning of a problem on a domain with unequal work estimates in different subdomains in a way that balances the workload across multiple processors. Such a problem arises, for example, in solving partial differential equations using an adaptive method that places extra grid points in certain subregions of the domain. A binary decomposition of the domain is used to partition it into rectangles requiring equal computational effort. The communication cost of mapping this partitioning onto different multiprocessors - is a mesh-connected array, a tree machine, and a hypercube - is then studied. The communication-cost expressions can be used to determine the optimal depth of the partitioning.

Original languageEnglish (US)
Pages (from-to)570-580
Number of pages11
JournalIEEE Transactions on Computers
VolumeC-36
Issue number5
StatePublished - May 1987

Fingerprint

Multiprocessor
Partitioning
Communication Cost
Communication
Partial differential equations
Costs
Adaptive Method
Decomposition
Hypercube
Unequal
Rectangle
Workload
Partial differential equation
Partition
Mesh
Binary
Grid
Decompose
Estimate
Strategy

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

PARTITIONING STRATEGY FOR NONUNIFORM PROBLEMS ON MULTIPROCESSORS. / Berger, Marsha; Bokhari, Shahid H.

In: IEEE Transactions on Computers, Vol. C-36, No. 5, 05.1987, p. 570-580.

Research output: Contribution to journalArticle

@article{1aff6fa40f98430e97d2c287fe5cd3ef,
title = "PARTITIONING STRATEGY FOR NONUNIFORM PROBLEMS ON MULTIPROCESSORS.",
abstract = "The authors consider the partitioning of a problem on a domain with unequal work estimates in different subdomains in a way that balances the workload across multiple processors. Such a problem arises, for example, in solving partial differential equations using an adaptive method that places extra grid points in certain subregions of the domain. A binary decomposition of the domain is used to partition it into rectangles requiring equal computational effort. The communication cost of mapping this partitioning onto different multiprocessors - is a mesh-connected array, a tree machine, and a hypercube - is then studied. The communication-cost expressions can be used to determine the optimal depth of the partitioning.",
author = "Marsha Berger and Bokhari, {Shahid H.}",
year = "1987",
month = "5",
language = "English (US)",
volume = "C-36",
pages = "570--580",
journal = "IEEE Transactions on Computers",
issn = "0018-9340",
publisher = "IEEE Computer Society",
number = "5",

}

TY - JOUR

T1 - PARTITIONING STRATEGY FOR NONUNIFORM PROBLEMS ON MULTIPROCESSORS.

AU - Berger, Marsha

AU - Bokhari, Shahid H.

PY - 1987/5

Y1 - 1987/5

N2 - The authors consider the partitioning of a problem on a domain with unequal work estimates in different subdomains in a way that balances the workload across multiple processors. Such a problem arises, for example, in solving partial differential equations using an adaptive method that places extra grid points in certain subregions of the domain. A binary decomposition of the domain is used to partition it into rectangles requiring equal computational effort. The communication cost of mapping this partitioning onto different multiprocessors - is a mesh-connected array, a tree machine, and a hypercube - is then studied. The communication-cost expressions can be used to determine the optimal depth of the partitioning.

AB - The authors consider the partitioning of a problem on a domain with unequal work estimates in different subdomains in a way that balances the workload across multiple processors. Such a problem arises, for example, in solving partial differential equations using an adaptive method that places extra grid points in certain subregions of the domain. A binary decomposition of the domain is used to partition it into rectangles requiring equal computational effort. The communication cost of mapping this partitioning onto different multiprocessors - is a mesh-connected array, a tree machine, and a hypercube - is then studied. The communication-cost expressions can be used to determine the optimal depth of the partitioning.

UR - http://www.scopus.com/inward/record.url?scp=0023346636&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0023346636&partnerID=8YFLogxK

M3 - Article

VL - C-36

SP - 570

EP - 580

JO - IEEE Transactions on Computers

JF - IEEE Transactions on Computers

SN - 0018-9340

IS - 5

ER -