Abstract
For efficient execution of parallel irregular computations, dynamic load balancing must be applied. If the computational work is associated with data sets, which must be separately processed by an algorithm, then load balancing can be performed most efficiently by transfering the data sets between processes using application level messages. Such a situation exists in parallel branch and bound (B&B) computations. A parallel B&B algorithm has been implemented in a novel parallel programming environment. This environment facilitates an infrastructure for parallel application control. Application consistent global states are continuously monitored. Control decisions are taken based on the monitored states and the decisions are communicated to the application processes. This infrastructure has been used for load balancing strategy implementation in parallel B&B computations. An analysis of the characteristics of the control infrastructure and the application resulted in a choice of a global load balancing strategy working with many simple and small steps executed frequently. Experiments have shown, that this strategy works well. The chosen strategy is much more efficient (shortening the application runtime by more than 3 times), if the prediction of the results of an already taken load balancing decision is used for subsequent load balancing decisions.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Wu, J.: Distributed System Design. CRC Press, Boca Raton (1999)
Zhou, S., Ferrari, D.: An experimental study of load balancing performance. Technical Report Technical Report UCB/CSD 87/336, University of California, Berkeley (1987)
Phillips, I.W., Capon, P.C.: Experiments in distributed load balancing. In: Boyanov, K. (ed.) Proceedings of the Fourth Workshop on Parallel and Distributed Processing, WP&DP 1993, Sofia, Bulgaria, pp. 146–171. North-Holland, Amsterdam (1993)
Iskra, K., Linden, F., der Hendrikse, Z.V., Overeinder, B., Albada, G., Sloot, B.V.: The implementation of dynamite: An environment for migrating PVM tasks. Operating Systems Review 34(3), 40–55 (2000)
Corradi, A., Leonardi, L., Zambonelli, F.: Dynamic load distribution in massively parallel architectures: The parallel objects example. In: 1st Conference on Massively Parallel Computing Systems, May 1994, pp. 318–322. IEEE, Los Alamitos (1994)
Philippsen, M., Zenger, M.: JavaParty - transparent remote objects in java. Concurrency: Practice and Experience 9(11), 1225–1242 (1997)
Schürer, R.: Optimal communication interval for parallel adaptive integration. Parallel and Distributed Computing Practices 5 (3), 269–278 (2002)
Bahi, J.M., Contassot-Vivier, S., Couturier, R.: Dynamic load balancing and efficient load estimators for asynchronous iterative algorithms. IEEE Transactions on Parallel and Distributed Systems 16(4), 289–299 (2005)
Czarnul, P., Kranzlmüller, D., Kacsuk, P., Dongarra, J., Volkert, J.: Development and tuning of irregular divide-and-conquer applications in DAMPVM/DAC. In: Kranzlmüller, D., Kacsuk, P., Dongarra, J., Volkert, J. (eds.) PVM/MPI 2002. LNCS, vol. 2474, pp. 208–216. Springer, Heidelberg (2002)
Stoller, S.D.: Detecting global predicates in distributed systems with clocks. Distributed Computing 13(2), 85–98 (2000)
Borkowski, J.: Global predicates for on-line control of distributed applications. In: Wyrzykowski, R., Dongarra, J., Paprzycki, M., Waśniewski, J. (eds.) PPAM 2004. LNCS, vol. 3019, pp. 269–277. Springer, Heidelberg (2004)
Borkowski, J., Kopanski, D., Tudruj, M.: Parallel irregular computations control based on global predicate monitoring. In: Proc. Of International Symposium on Parallel Computing in Electrical Engineering PARELEC 2006, Bialystok, Poland, pp. 233–238. IEEE, Los Alamitos (2006)
Borkowski, J., Tudruj, M., Kopanski, D.: Parallel program design tool with application control methods based on global states. In: Wyrzykowski, R., Dongarra, J., Paprzycki, M., Waśniewski, J. (eds.) PPAM 2004. LNCS, vol. 3019, pp. 338–343. Springer, Heidelberg (2004)
Borkowski, J.: Parallel program control based on hierarchically detected consistent global states. In: International Conference on Parallel Computing in Electrical Engineering (PARELEC 2004), pp. 328–333. IEEE, Los Alamitos (2004)
Biswas, R., Oliker, L., Shan, H.: Parallel computing strategies for irregular algorithms. Annual Review of Scalable Computing (April 2003)
Androulakis, I.P., Floudas, C.A.: Distributed branch and bound algorithms for global optimization. The IMA Volumes in Mathemathics and its Applications, ”Parallel Processig of Discreet Problems” 106, 1–37 (1999)
Borkowski, J., Tudruj, M.: Dual communication network in program control based on global application state monitoring. In: Proceedings of ISPDC 2007, Linz, Austria, IEEE, Los Alamitos (accepted for publication, 2007)
Eager, D., Lazowska, E., Zahorjan, J.: Adaptive load sharing in homogeneous distributed systems. IEEE Transactions on Software Engineering, 662–675 (May 1986)
Trienekens, H.: Parallel branch and bound and anomalies. Technical Report EUR-FEW-CS-89-01, Erasmus University Rotterdam (1989)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Borkowski, J., Tudruj, M. (2008). Parallel Irregular Computations with Dynamic Load Balancing through Global Consistent State Monitoring. In: Wyrzykowski, R., Dongarra, J., Karczewski, K., Wasniewski, J. (eds) Parallel Processing and Applied Mathematics. PPAM 2007. Lecture Notes in Computer Science, vol 4967. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-68111-3_103
Download citation
DOI: https://doi.org/10.1007/978-3-540-68111-3_103
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-68105-2
Online ISBN: 978-3-540-68111-3
eBook Packages: Computer ScienceComputer Science (R0)