skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Parallel hypergraph partitioning for irregular problems.

Abstract

No abstract prepared.

Authors:
; ;  [1];  [2];
  1. (Utrecht University, The Netherlands)
  2. (Ohio State University, Columbus)
Publication Date:
Research Org.:
Sandia National Laboratories
Sponsoring Org.:
USDOE
OSTI Identifier:
943935
Report Number(s):
SAND2006-1083C
TRN: US200902%%382
DOE Contract Number:
AC04-94AL85000
Resource Type:
Conference
Resource Relation:
Conference: Proposed for presentation at the SIAM 2006 Conference on Pallel Processing held February 22-24, 2006 in San Francisco, CA.
Country of Publication:
United States
Language:
English
Subject:
97; 99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; PARALLEL PROCESSING; DIAGRAMS; PARTITION FUNCTIONS

Citation Formats

Heaphy, Robert, Devine, Karen Dragon, Bisseling, Rob, Catalyurek, Umit, and Boman, Erik Gunnar. Parallel hypergraph partitioning for irregular problems.. United States: N. p., 2006. Web.
Heaphy, Robert, Devine, Karen Dragon, Bisseling, Rob, Catalyurek, Umit, & Boman, Erik Gunnar. Parallel hypergraph partitioning for irregular problems.. United States.
Heaphy, Robert, Devine, Karen Dragon, Bisseling, Rob, Catalyurek, Umit, and Boman, Erik Gunnar. Wed . "Parallel hypergraph partitioning for irregular problems.". United States. doi:.
@article{osti_943935,
title = {Parallel hypergraph partitioning for irregular problems.},
author = {Heaphy, Robert and Devine, Karen Dragon and Bisseling, Rob and Catalyurek, Umit and Boman, Erik Gunnar},
abstractNote = {No abstract prepared.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Wed Feb 01 00:00:00 EST 2006},
month = {Wed Feb 01 00:00:00 EST 2006}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • Graph partitioning is often used for load balancing in parallel computing, but it is known that hypergraph partitioning has several advantages. First, hypergraphs more accurately model communication volume, and second, they are more expressive and can better represent nonsymmetric problems. Hypergraph partitioning is particularly suited to parallel sparse matrix-vector multiplication, a common kernel in scientific computing. We present a parallel software package for hypergraph (and sparse matrix) partitioning developed at Sandia National Labs. The algorithm is a variation on multilevel partitioning. Our parallel implementation is novel in that it uses a two-dimensional data distribution among processors. We present empirical resultsmore » that show our parallel implementation achieves good speedup on several large problems (up to 33 million nonzeros) with up to 64 processors on a Linux cluster.« less
  • Abstract not provided.
  • Regular distributions for storing dense matrices on parallel systems are not always used in practice. In many scientific applications, matrix distribution is based on the underlying physical problem and might involve variable block sizes on individual processors. This paper describes a generalization of the Shared and Remote-memory based Universal Matrix Multiplication Algorithm (SRUMMA) [1] to handle irregularly distributed matrices. Our approach relies on a distribution independent algorithm that provides dynamic load balancing by exploiting data locality and achieves performance as good as the traditional approach which relies on temporary arrays with regular distribution, data redistribution, and matrix multiplication for regularmore » matrices to handle the irregular case. The proposed algorithm is memory-efficient because temporary matrices are not needed. This feature is critical for systems like the IBM Blue Gene/L that offer very limited amount of memory per node. The experimental results demonstrate very good performance across the range matrix distributions and problem sizes motivated by real applications.« less