Parallelizing a Skyline Matrix Solver using Orca

Comparing parallel or distributed programming systems is, due to the variety of available flavors of paradigms, methods and solutions, not a task to be taken lightly. The Cowichan problem set aims to provide a collection of non-trivial problems of which the implementation can greatly help in assessi...

Full description

Bibliographic Details
Main Author: David S. Bouman
Other Authors: The Pennsylvania State University CiteSeerX Archives
Format: Text
Language:English
Published: 1995
Subjects:
Online Access:http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.31.436
Description
Summary:Comparing parallel or distributed programming systems is, due to the variety of available flavors of paradigms, methods and solutions, not a task to be taken lightly. The Cowichan problem set aims to provide a collection of non-trivial problems of which the implementation can greatly help in assessing the usability of a particular system. The skyline matrix problem of this set can be used to test the ease with which a system is able to handle irregular matrix structures. This paper reports the development, experiences and results of the implementation of this problem using the Orca parallel programming language. 1 1 Introduction Parallelism, if it can be exploited, promises to offer much higher performance at a fraction of the cost compared to traditional sequential systems. If it can be exploited that is. The problem in many cases is that creating parallel programs is not at all a trivial task. Programming parallel systems usually implies more things to keep track of by the p.