Computing the Throughput of Probabilistic and Replicated Streaming Applications
- 179 Downloads
In this paper, we investigate how to compute the throughput of probabilistic and replicated streaming applications. We are given (i) a streaming application whose dependence graph is a linear chain; (ii) a one-to-many mapping of the application onto a fully heterogeneous target platform, where a processor is assigned at most one application stage, but where a stage can be replicated onto a set of processors; and (iii) a set of random variables modeling the computation and communication times in the mapping. We show how to compute the throughput of the application, i.e., the rate at which data sets can be processed. The problem is easy when application stages are not replicated, i.e., each application stage is assigned to a single processor: in that case the throughput is dictated by the critical hardware resource. However, when stages are replicated, i.e., each application stage may be assigned to several processors, the problem becomes surprisingly complicated: even in the deterministic case, the optimal throughput may be lower than the smallest internal resource throughput.
The first contribution of the paper is to provide a general method to compute the throughput when computation and communication times, also called stage parameters, are constant or follow I.I.D. exponential laws. The second contribution is to provide bounds for the throughput when stage parameters form associated random sequences (correlation between communication and processing times of a given data set on the different application stages, i.e., a data set that takes a long time on the first stage is likely to be large, and to take a long time on the next stages), and are N.B.U.E. (New Better than Used in Expectation) variables (if an operation has already been processed for some duration, the remaining time is smaller than the processing time of a fresh operation): the throughput is bounded from below by the exponential case and bounded from above by the deterministic case. An extensive set of simulation allows us to assess the quality of the model, and to observe the actual behavior of several distributions.
KeywordsCommunication Time Deterministic Case Critical Resource Overlap Model Streaming Application
The authors thank the reviewers for their numerous comments and suggestions, which greatly improved the final version of the paper.
- 4.Benoit, A., Gallet, M., Gaujal, B., Robert, Y.: Computing the throughput of replicated workflows on heterogeneous platforms. In: Proceedings of ICPP’2009, the 38th International Conference on Parallel Processing (2009) Google Scholar
- 5.Benoit, A., Dufossé, F., Gallet, M., Gaujal, B., Robert, Y.: Computing the throughput of probabilistic and replicated streaming applications. In: Proceedings of SPAA 2010, the 22nd ACM Symposium on Parallelism in Algorithms and Architectures. ACM, New York (2010) Google Scholar
- 10.DataCutter Project. Middleware for filtering large archival scientific datasets in a grid environment. http://www.cs.umd.edu/projects/hpsl/ResearchAreas/DataCutter.htm
- 12.Gaujal, B., Vincent, J.-M.: Comparisons of stochastic task-resource systems. In: Introduction to Scheduling. CRC Press, Boca Raton (2009) Google Scholar
- 15.Jean-Marie, A.: ERS: a tool set for performance evaluation of discrete event systems. http://www-sop.inria.fr/mistral/soft/ers.html
- 17.Knuth, D.E.: The Art of Computer Programming, vol. 3, 2nd edn. Addison-Wesley, Reading (1998) Google Scholar
- 19.Spencer, M., Ferreira, R., Beynon, M., Kurc, T., Catalyurek, U., Sussman, A., Saltz, J.: Executing multiple pipelined data analysis operations in the grid. In: Proceedings of Supercomputing’02, the 2002 ACM/IEEE Conference on Supercomputing, pp. 1–18. IEEE Comput. Soc., Los Alamitos (2002) Google Scholar
- 20.Subhlok, J., Vondran, G.: Optimal mapping of sequences of data parallel tasks. In: Proceedings of PPoPP’95, the ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, pp. 134–143. ACM, New York (1995) Google Scholar
- 21.Subhlok, J., Vondran, G.: Optimal latency-throughput tradeoffs for data parallel pipelines. In: Proceedings of SPAA’96, the 8th ACM Symposium on Parallel Algorithms and Architectures, pp. 62–71. ACM, New York (1996) Google Scholar
- 22.Taura, K., Chien, A.: A heuristic algorithm for mapping communicating tasks on heterogeneous resources. In: Proceedings of HCW’00, the 9th Heterogeneous Computing Workshop, pp. 102–115. IEEE Comput. Soc., Los Alamitos (2000) Google Scholar
- 23.Velho, P., Legrand, A.: Accuracy study and improvement of network simulation in the SimGrid framework. In: Proceedings of Simutools’09, the 2nd International Conference on Simulation Tools and Techniques, ICST, pp. 1–10 (2009) Google Scholar
- 24.Vydyanathan, N., Çatalyurek, Ü.V., Kurc, T., Saddayappan, P., Saltz, J.: Toward optimizing latency under throughput constraints for application workflows on clusters. In: Proceedings of Euro-Par’07. LNCS, vol. 4641, pp. 173–183. Springer, Berlin (2007) Google Scholar
- 25.Vydyanathan, N., Çatalyurek, Ü.V., Kurc, T., Saddayappan, P., Saltz, J.: A duplication based algorithm for optimizing latency under throughput constraints for streaming workflows. In: Proceedings of ICPP’2008, the 37th International Conference on Parallel Processing, pp. 254–261. IEEE Comput. Soc., Los Alamitos (2008) Google Scholar