Static scheduling using task replication for LogP and BSP models
This paper investigates the effect of communication overhead on the scheduling problem. We present a scheduling algorithm, based on LogP-type models, for allocating task graphs to networks of processors. The makespans of schedules produced by our multi-stage scheduling approach (MSA) are compared with other well-known scheduling heuristics. The results indicate that new classes of scheduling heuristics are required to generate efficient schedules for realistic abstractions of today’s parallel computers. The scheduling strategy of MSA can also be used to generate BSP-structured programs from more abstract representations. The performance of such programs are compared with “conventional” versions.
Unable to display preview. Download preview PDF.
- 1.C. Boeres. Versatile Communication Cost Modelling for Multicomputer Task Scheduling Heuristics. PhD thesis, Department of Computer Science, University of Edinburgh, May 1997.Google Scholar
- 2.C. Boeres and V. E. F. Rebello. Versatile task scheduling of binary trees for realistic machines. In M. Griebl C. Lengauer and S. Gorlatch, editors, Proceedings of the 3rd International Euro-Par Conference on Parallel Processing (Euro-Par’97), LNCS 1300, pages 913–921, Passau, Germany, August 1997. Springer-Verlag.Google Scholar
- 3.C. Boeres and V. E. F. Rebello. A versatile cost modelling approach for multicomputer task scheduling. Accepted for publication in Parallel Computing, 1998.Google Scholar
- 5.G. Chochia, C. Boeres, M. Norman, and P. Thanisch. Analysis of multicomputer schedules in a cost and latency model of communication. In Proceedings of the 3rd Workshop on Abstract Machine Models for Parallel and Distributed Computing, Leeds, UK., April 1996. IOS press.Google Scholar
- 6.D. Culler et al. LogP: Towards a realistic model of parallel computation. In Proceedings of the Fourth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, San Diego, CA, May 1993.Google Scholar
- 7.B. Di Martino and G. Ianello. Parallelization on non-simultaneous iterative methods for systems of linear equations. In Parallel Processing (CONPAR’94-VAPP VI), LNCS 854, pages 254–264. Springer-Verlag, 1994.Google Scholar
- 8.J. M. D. Hill, P. I. Crumpton, and D. A. Burgess. The theory, practice, and a tool for BSP performance prediction. In Proceedings of the 2nd International Euro-Par Conference on Parallel Processing (Euro-Par’96), volume 1124 of LNCS, pages 697–705. Springer-Verlag, August 1996.Google Scholar
- 9.J. M. D. Hill, S. Donaldson, and D. B. Skillicorn. Stability of communication performance in practice: from the Cray T3E to networks of workstations. Technical Report PRG-TR-33-97, Oxford University Computing Laboratory, October 1997.Google Scholar
- 12.H. Jung, L. Kirousis, and P. Spirakis. Lower bounds and efficient algorithms for multiprocessor scheduling of DAGs with communication delays. In Proc. ACM Symposium on Parallel Algorithms and Architectures, pages 254–264, 1989.Google Scholar
- 13.W. Lowe, W. Zimmermann, and J. Eisenbiergler. Optimization of parallel programs on machines with expensive communication. In L. Bouge, P. Fraigniaud, A. Mignotte, and Y. Robert, editors, Proceedings of the 2nd International Euro-Par Conference on Parallel Processing (Euro-Par’96), LNCS 1124, pages 602–610, Lyon, France, August 1996. Springer-Verlag.Google Scholar
- 16.D. B. Skillicorn, J. M. D. Hill, and W. F. McColl. Questions and answers about BSP. Scientific Programming, 6(3):249–274, 1997.Google Scholar