Developing non-local iterative parallel algorithms for GIS on Cray T3D using MPI

A Clematis, A Coda, M Spagnuolo, M Mineter, Terence Sloan

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

Abstract

In the application context of spatial data analysis, several parallelisation strategies are proposed for algorithms which are characterized by the non-locality of computation and by the iterative nature of the data processing. Different parallelisation approaches are presented and the methods developed for Cray T3D using the MPI message passing library are briefly compared to a workstation-network solution. The better results in the two environments are obtained following different strategies. The lesson learned is that MPI and other message passing libraries provide valuable help for portability, but the development of efficient and portable parallel software needs the availability of higher level approaches supporting different parallelisation strategies in a simple and effective way.
Original languageEnglish
Title of host publicationRecent Advances in Parallel Virtual Machine and Message Passing Interface
Subtitle of host publication4th European PVM/MPI User's Group Meeting Cracow, Poland, November 3-5, 1997, Proceedings
EditorsMarian Bubak, Jack Dongarra, Jerzy Wasniewski
PublisherSpringer
Pages435
Number of pages8
Volume1332
ISBN (Print)978-3-540-63697-7
DOIs
Publication statusPublished - 1997

Publication series

NameLecture Notes in Computer Science,

Fingerprint

Dive into the research topics of 'Developing non-local iterative parallel algorithms for GIS on Cray T3D using MPI'. Together they form a unique fingerprint.

Cite this