Managing the Bottlenecks in Parallel Gauss-Seidel Type Algorithms for Power Flow Analysis

Garng Morton Huang, W. Ongsakul

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

In our earlier papers, the parallelization and implementations of Gauss-Seidel (G-S) algorithms for power flow analysis have been investigated on a Sequent Balance shared memory (SM) machine. In this paper, we generalize the idea to more general computer architectures and demonstrate how to effectively increase the speedup upper bounds of G-S algorithms by properly managing the bottlenecks on both Sequent Balance SM and nCUBE2 distributed memory (DM) machines. For G-S algorithms, when our coloring process is used to schedule the processors, there is almost no sequential portion. Thus, the only decisive factor left, which has a direct impact on the speedup upper bound, is the synchronization overhead. Accordingly, we propose a new synchronization scheme which can reduce the synchronization overhead on the Sequent Balance machine. Also, on the CUBE2 machine, a new clustered G-S algorithm is proposed and implemented. The algorithm carefully schedules their processors, computational loads and communication overheads for the best performance. In addition, the synchronization overheads and speedup upper bounds on both machines are analyzed in terms of power system size and number of processors. The competitiveness of G-S type algorithms is also discussed.

Original languageEnglish
Pages (from-to)677-684
Number of pages8
JournalIEEE Transactions on Power Systems
Volume9
Issue number2
DOIs
Publication statusPublished - 1 Jan 1994
Externally publishedYes

Fingerprint

Synchronization
Data storage equipment
Computer architecture
Coloring
Communication

Keywords

  • CUBE2 DM machine
  • Gauss-Seidel
  • Sequent Balance SM machine

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering

Cite this

Managing the Bottlenecks in Parallel Gauss-Seidel Type Algorithms for Power Flow Analysis. / Huang, Garng Morton; Ongsakul, W.

In: IEEE Transactions on Power Systems, Vol. 9, No. 2, 01.01.1994, p. 677-684.

Research output: Contribution to journalArticle

@article{c688292b12a746a8aad29127078e7c23,
title = "Managing the Bottlenecks in Parallel Gauss-Seidel Type Algorithms for Power Flow Analysis",
abstract = "In our earlier papers, the parallelization and implementations of Gauss-Seidel (G-S) algorithms for power flow analysis have been investigated on a Sequent Balance shared memory (SM) machine. In this paper, we generalize the idea to more general computer architectures and demonstrate how to effectively increase the speedup upper bounds of G-S algorithms by properly managing the bottlenecks on both Sequent Balance SM and nCUBE2 distributed memory (DM) machines. For G-S algorithms, when our coloring process is used to schedule the processors, there is almost no sequential portion. Thus, the only decisive factor left, which has a direct impact on the speedup upper bound, is the synchronization overhead. Accordingly, we propose a new synchronization scheme which can reduce the synchronization overhead on the Sequent Balance machine. Also, on the CUBE2 machine, a new clustered G-S algorithm is proposed and implemented. The algorithm carefully schedules their processors, computational loads and communication overheads for the best performance. In addition, the synchronization overheads and speedup upper bounds on both machines are analyzed in terms of power system size and number of processors. The competitiveness of G-S type algorithms is also discussed.",
keywords = "CUBE2 DM machine, Gauss-Seidel, Sequent Balance SM machine",
author = "Huang, {Garng Morton} and W. Ongsakul",
year = "1994",
month = "1",
day = "1",
doi = "10.1109/59.317675",
language = "English",
volume = "9",
pages = "677--684",
journal = "IEEE Transactions on Power Systems",
issn = "0885-8950",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

TY - JOUR

T1 - Managing the Bottlenecks in Parallel Gauss-Seidel Type Algorithms for Power Flow Analysis

AU - Huang, Garng Morton

AU - Ongsakul, W.

PY - 1994/1/1

Y1 - 1994/1/1

N2 - In our earlier papers, the parallelization and implementations of Gauss-Seidel (G-S) algorithms for power flow analysis have been investigated on a Sequent Balance shared memory (SM) machine. In this paper, we generalize the idea to more general computer architectures and demonstrate how to effectively increase the speedup upper bounds of G-S algorithms by properly managing the bottlenecks on both Sequent Balance SM and nCUBE2 distributed memory (DM) machines. For G-S algorithms, when our coloring process is used to schedule the processors, there is almost no sequential portion. Thus, the only decisive factor left, which has a direct impact on the speedup upper bound, is the synchronization overhead. Accordingly, we propose a new synchronization scheme which can reduce the synchronization overhead on the Sequent Balance machine. Also, on the CUBE2 machine, a new clustered G-S algorithm is proposed and implemented. The algorithm carefully schedules their processors, computational loads and communication overheads for the best performance. In addition, the synchronization overheads and speedup upper bounds on both machines are analyzed in terms of power system size and number of processors. The competitiveness of G-S type algorithms is also discussed.

AB - In our earlier papers, the parallelization and implementations of Gauss-Seidel (G-S) algorithms for power flow analysis have been investigated on a Sequent Balance shared memory (SM) machine. In this paper, we generalize the idea to more general computer architectures and demonstrate how to effectively increase the speedup upper bounds of G-S algorithms by properly managing the bottlenecks on both Sequent Balance SM and nCUBE2 distributed memory (DM) machines. For G-S algorithms, when our coloring process is used to schedule the processors, there is almost no sequential portion. Thus, the only decisive factor left, which has a direct impact on the speedup upper bound, is the synchronization overhead. Accordingly, we propose a new synchronization scheme which can reduce the synchronization overhead on the Sequent Balance machine. Also, on the CUBE2 machine, a new clustered G-S algorithm is proposed and implemented. The algorithm carefully schedules their processors, computational loads and communication overheads for the best performance. In addition, the synchronization overheads and speedup upper bounds on both machines are analyzed in terms of power system size and number of processors. The competitiveness of G-S type algorithms is also discussed.

KW - CUBE2 DM machine

KW - Gauss-Seidel

KW - Sequent Balance SM machine

UR - http://www.scopus.com/inward/record.url?scp=0028427679&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028427679&partnerID=8YFLogxK

U2 - 10.1109/59.317675

DO - 10.1109/59.317675

M3 - Article

VL - 9

SP - 677

EP - 684

JO - IEEE Transactions on Power Systems

JF - IEEE Transactions on Power Systems

SN - 0885-8950

IS - 2

ER -