Abstract
Parallel/distributed application development is an extremely difficult task for non-expert programmers, and support tools are therefore needed for all phases of the development cycle of this kind of applications. In particular, dynamic performance tuning tools can take advantage of the knowledge about the application’s structure given by a skeleton based programming tool. This study shows the definition of a strategy for dynamically improving the performance of pipeline applications. This strategy, which has been called Dynamic Pipeline Mapping, improves the application’s throughput by gathering the pipe’s fastest stages and replicating its slowest ones. We have evaluated the new algorithm by experimentation and simulation, and results show that DPM leads to significant performance improvements.
This work was supported by MEC under contract TIN2007-64974.
Chapter PDF
Similar content being viewed by others
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
References
Cesar, E., Moreno, A., Sorribes, J., Luque, E.: Modeling master/worker applications for automatic performance tuning. Parallel Comput. 32(7), 568–589 (2006)
Cesar, E., Sorribes, J., Luque, E.: Modeling pipeline applications in poetries. In: Cunha, J.C., Medeiros, P.D. (eds.) Euro-Par 2005. LNCS, vol. 3648, pp. 83–95. Springer, Heidelberg (2005)
Morajko, A., Cesar, E., Caymes-Scutari, P., Margalef, T., Sorribes, J., Luque, E.: Automatic tuning of master/worker applications. In: Cunha, J.C., Medeiros, P.D. (eds.) Euro-Par 2005. LNCS, vol. 3648, pp. 95–103. Springer, Heidelberg (2005)
Subhlok, J., Vondran, G.: Optimal use of mixed task and data parallelism for pipelined computations. Journal of Parallel and Distributed Computing 60(3), 297–319 (2000)
Lee, M., Liu, W., Prasanna, V.K.: A mapping methodology for designing software task pipelines for embedded signal processing. In: IPPS/SPDP Workshops, pp. 937–944 (1998)
Hoang, P.D., Rabaey, J.: Scheduling of dsp programs onto multiprocessors for maximum throughput. IEEE Transactions on Signal Processing 41(6), 2225–2235 (1993)
Yang, M.-T., Kasturi, R., Sivasubramaniam, A.: A pipeline-based approach for scheduling video processing algorithms on now. Transactions on Parallel and Distributed Systems 14(2), 119–130 (2003)
Guirado, F., Ripoll, A., Roig, C., Luque, E.: Exploitation of parallelism for applications with an input data stream: optimal resource-throughput tradeoffs. In: PDP 2005. 13th Euromicro Conference on Parallel, Distributed and Network-Based Processing, 9-11 February, pp. 170–178 (2005)
Almeida, F., González, D., Moreno, L.M., Rodríguez, C.: An analytical model for pipeline algorithms on heterogeneous clusters. In: Proceedings of the 9th European PVM/MPI Users’ Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface, London, UK, pp. 441–449. Springer, Heidelberg (2002)
Gilmore, S., Hillston, J., Kloul, L., Ribaudo, M.: Pepa nets: a structured performance modelling formalism. Perform. Eval. 54(2), 79–104 (2003)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Moreno, A., César, E., Guevara, A., Sorribes, J., Margalef, T., Luque, E. (2008). Dynamic Pipeline Mapping (DPM). In: Luque, E., Margalef, T., Benítez, D. (eds) Euro-Par 2008 – Parallel Processing. Euro-Par 2008. Lecture Notes in Computer Science, vol 5168. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-85451-7_32
Download citation
DOI: https://doi.org/10.1007/978-3-540-85451-7_32
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-85450-0
Online ISBN: 978-3-540-85451-7
eBook Packages: Computer ScienceComputer Science (R0)