Abstract
Parallel programs experience performance inefficiencies as a result of dependencies, resource contentions, uneven work distributions and loss of synchronizations among processors. The analysis of these inefficiencies is very important for tuning and performance debugging studies. In this paper we address the identification and localization of performance inefficiencies from a methodological viewpoint. We follow a top down approach. We first analyze the performance properties of the programs at a coarse grain. We then study the behavior of the processors and their load imbalance. The methodology is illustrated on a study of a message passing computational fluid dynamic program.
This work has been supported by the Italian Ministry of Education, University and Research (MIUR) under the FIRB Programme, by the University of Pavia under the FAR Programme and by the Italian Research Council (CNR).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Calzarossa, M., Massari, L., Merlo, A., Pantano, M., Tessera, D.: Medea: A Tool for Workload Characterization of Parallel Systems. IEEE Parallel and Distributed Technology 3(4), 72–80 (1995)
De Rose, L., Zhang, Y., Reed, D.A.: SvPablo: A Multi-Language Performance Analysis System. In: Puigjaner, R., Savino, N.N., Serra, B. (eds.) TOOLS 1998. LNCS, vol. 1469, pp. 352–355. Springer, Heidelberg (1998)
Ferschweiler, K., Harrah, S., Keon, D., Calzarossa, M., Tessera, D., Pancake, C.: The Tracefile Testbed – A Community Repository for Identifying and Retrieving HPC Performance Data. In: Proc. 2002 International Conference on Parallel Processing, pp. 177–184. IEEE Press, Los Alamitos (2002)
Hartigan, J.A.: Clustering Algorithms. Wiley, Chichester (1975)
Heath, M.T., Etheridge, J.A.: Visualizing the Performance of Parallel Programs. IEEE Software 8, 29–39 (1991)
Helm, B., Malony, A., Fickas, S.: Capturing and Automating Performance Diagnosis: the Poirot Approach. In: Proceedings of the 1995 International Parallel Processing Symposium, pp. 606–613 (1995)
Karavanic, K.L., Miller, B.P.: Improving Online Performance Diagnosis by the Use of Historical Performance Data. In: Proc. SC 1999 (1999)
Marshall, A.W., Olkin, I.: Inequalities: Theory of Majorization and Its Applications. Academic Press, London (1979)
Miller, B.P., Callaghan, M.D., Cargille, J.M., Hollingsworth, J.K.H., Irvin, R.B., Karavanic, K.L., Kunchithapadam, K., Newhall, T.: The Paradyn Parallel Measurement Performance Tool. IEEE Computer 28(11), 37–46 (1995)
Roth, P.C., Miller, B.P.: Deep Start: A Hybrid Strategy for Automated Performance Problem Searches. In: Monien, B., Feldmann, R.L. (eds.) Euro-Par 2002. LNCS, vol. 2400, pp. 86–96. Springer, Heidelberg (2002)
Simmons, M.L., Hayes, A.H., Brown, J.S., Reed, D.A. (eds.): Debugging and Performance Tuning for Parallel Computing Systems. IEEE Computer Society, Los Alamitos (1996)
Williams, W., Hoel, T., Pase, D.: The MPP Apprentice Performance Tool: Delivering the Performance of the Cray T3D. In: Decker, K.M. (ed.) Programming Environments for Massively Parallel Distributed Systems, pp. 333–345. Birkhauser, Basel (1994)
Yan, J.C., Sarukkai, S.R.: Analyzing Parallel Program Performance Using Normalized Performance Indices and Trace Transformation Techniques. Parallel Computing 22(9), 1215–1237 (1996)
Zaki, O., Lusk, E., Gropp, W., Swider, D.: Toward Scalable Performance Visualization with Jumpshot. The International Journal of High Performance Computing Applications 13(2), 277–288 (1999)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Calzarossa, M., Massari, L., Tessera, D. (2003). Load Imbalance in Parallel Programs. In: Malyshkin, V.E. (eds) Parallel Computing Technologies. PaCT 2003. Lecture Notes in Computer Science, vol 2763. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-45145-7_17
Download citation
DOI: https://doi.org/10.1007/978-3-540-45145-7_17
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-40673-0
Online ISBN: 978-3-540-45145-7
eBook Packages: Springer Book Archive