Abstract
This paper presents different schemes to parallelize a code that solves a Network Flow Optimization problem. These schemes have been implemented on a heterogeneous cluster of workstations with a multiprogrammed workload. The paper shows that for this situation, it is crucial to have a load balancing strategy that takes into account the computing power of each node as well as its load. Performance figures show that significant speed-ups are obtained when this load balancing strategy is followed.
Preview
Unable to display preview. Download preview PDF.
References
Labarta, J., Girona, S., Pillet, V., Cortes, T., Cela, J.M.: A Parallel Program Developement Environment: CEPBA/UPC Report No. RR-95/02 (1995)
Ahuja, R., Magnanti, T., Orlin, J.: Network Flows, Theory, Algorithms and Applications: Prentice Hall, (1993)
Dembo, R.S.: Scenario optimization: Annals of Operation Research 30 (1991) 43–72.
Gueist, A., Beguelin, A., Dongarra, J., Jiang, W., Manchel, R., Sunderam, V.: PVM User's Guide and Reference Manual: Oak Ridge National Laboratory, TM-12187 (May 1994)
Manchek, R.: Design and Implementation of PVM 3 M.S. Degree, Thesis, University of Tennessee (May 1994)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Fontdecaba, E., González, A., Labarta, J. (1996). Load balancing in a Network Flow Optimization code. In: Dongarra, J., Madsen, K., Waśniewski, J. (eds) Applied Parallel Computing Computations in Physics, Chemistry and Engineering Science. PARA 1995. Lecture Notes in Computer Science, vol 1041. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-60902-4_25
Download citation
DOI: https://doi.org/10.1007/3-540-60902-4_25
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-60902-5
Online ISBN: 978-3-540-49670-0
eBook Packages: Springer Book Archive