Parallel solution algorithms for the triangular sylvester equation

A. Gerasoulis, I. Nelken

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The parallelization problem can be divided into three main stages: identification of parallelism which includes dependency analysis, partitioning the statements into atomic tasks of granularity suitable to the target architecture and scheduling these tasks into the processors. An MIMD coarse grained parallel algorithm is developed for the triangular Sylvester equation. We compare well known scheduling heuristics such as the naive and compute-ahead with the N-cp/misf methods which are described here. These methods trade off time and space according to the value of the parameter N . Our conclusion is that the N-cp/misf methods are faster than compute-ahead.

Original languageEnglish (US)
Title of host publicationApplications
EditorsDavid W. Walker, Quentin F. Stout
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages291-300
Number of pages10
ISBN (Electronic)0818621133, 9780818621130
DOIs
StatePublished - Jan 1 1990
Event5th Distributed Memory Computing Conference, DMCC 1990 - Charleston, United States
Duration: Apr 8 1990Apr 12 1990

Publication series

NameProceedings of the 5th Distributed Memory Computing Conference, DMCC 1990
Volume1

Conference

Conference5th Distributed Memory Computing Conference, DMCC 1990
Country/TerritoryUnited States
CityCharleston
Period4/8/904/12/90

ASJC Scopus subject areas

  • Hardware and Architecture
  • Information Systems and Management
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Parallel solution algorithms for the triangular sylvester equation'. Together they form a unique fingerprint.

Cite this