OPTIMIZING SKELETAL STREAM PROCESSING FOR DIVIDE AND CONQUER

Michael Poldner, Herbert Kuchen

2008

Abstract

Algorithmic skeletons intend to simplify parallel programming by providing recurring forms of program structure as predefined components. We present a new distributed task parallel skeleton for a very general class of divide and conquer algorithms for MIMD machines with distributed memory. Our approach combines skeletal internal task parallelism with stream parallelism. This approach is compared to alternative topologies for a task parallel divide and conquer skeleton with respect to their aptitude of solving streams of divide and conquer problems. Based on experimental results for matrix chain multiplication problems, we show that our new approach enables a better processor load and memory utilization of the engaged solvers, and reduces communication costs.

Paper Citation

in Harvard Style

Poldner M. and Kuchen H. (2008). OPTIMIZING SKELETAL STREAM PROCESSING FOR DIVIDE AND CONQUER . In Proceedings of the Third International Conference on Software and Data Technologies - Volume 1: ICSOFT, ISBN 978-989-8111-51-7, pages 181-189. DOI: 10.5220/0001889301810189

in Bibtex Style

@conference{icsoft08,
author={Michael Poldner and Herbert Kuchen},
title={OPTIMIZING SKELETAL STREAM PROCESSING FOR DIVIDE AND CONQUER},
booktitle={Proceedings of the Third International Conference on Software and Data Technologies - Volume 1: ICSOFT,},
year={2008},
pages={181-189},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001889301810189},
isbn={978-989-8111-51-7},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Third International Conference on Software and Data Technologies - Volume 1: ICSOFT,
TI - OPTIMIZING SKELETAL STREAM PROCESSING FOR DIVIDE AND CONQUER
SN - 978-989-8111-51-7
AU - Poldner M.
AU - Kuchen H.
PY - 2008
SP - 181
EP - 189
DO - 10.5220/0001889301810189