Productive Cluster Programming with OmpSs
Clusters of SMPs are ubiquitous. They have been traditionally programmed by using MPI. But, the productivity of MPI programmers is low because of the complexity of expressing parallelism and communication, and the difficulty of debugging. To try to ease the burden on the programmer new programming models have tried to give the illusion of a global shared-address space (e.g., UPC, Co-array Fortran). Unfortunately, these models do not support, increasingly common, irregular forms of parallelism that require asynchronous task parallelism. Other models, such as X10 or Chapel, provide this asynchronous parallelism but the programmer is required to rewrite entirely his application.
We present the implementation of OmpSs for clusters, a variant of OpenMP extended to support asynchrony, heterogeneity and data movement for task parallelism. As OpenMP, it is based on decorating an existing serial version with compiler directives that are translated into calls to a runtime system that manages the parallelism extraction and data coherence and movement. Thus, the same program written in OmpSs can run in a regular SMP machine, in clusters of SMPs, or even can be used for debugging with the serial version. The runtime uses the information provided by the programmer to distribute the work across the cluster while optimizes communications using affinity scheduling and caching of data.
We have evaluated our proposal with a set of kernels and the OmpSs versions obtain a performance comparable, or even superior, to the one obtained by the same version of MPI.
Unable to display preview. Download preview PDF.
- 1.Ayguade, E., Badia, R., Cabrera, D., Duran, A., Gonzalez, M., Igual, F., Jimenez, D., Labarta, J., Martorell, X., Mayo, R., Perez, J.M., Quintana-Orti, E.S.: A Proposal to Extend the OpenMP Tasking Model for Heterogeneous Architectures. In: IWOMP: Evolving OpenMP in an Age of Extreme Parallelism, Dresden, Germany, pp. 154–167 (June 2009)Google Scholar
- 2.Basumallik, A., Eigenmann, R.: Towards automatic translation of openmp to mpi. In: Proceedings of the 19th Annual International Conference on Supercomputing, ICS 2005, pp. 189–198. ACM, New York (2005)Google Scholar
- 4.Bonachea, D.: GASNet Specification, v1.8. Technical report, U.C. Berkeley (2006)Google Scholar
- 6.Charles, P., Grothoff, C., Saraswat, V., Donawa, C., Kielstra, A., Ebcioglu, K., von Praun, C., Sarkar, V.: X10: an object-oriented approach to non-uniform cluster computing. In: Proceedings of the 20th Annual ACM SIGPLAN Conference on Object-Oriented Programming, Systems, Languages, and Applications, OOPSLA 2005, New York, NY, USA (2005)Google Scholar
- 7.UPC Consortium. UPC Language Specifications v1.2 (May 2005)Google Scholar
- 8.Costa, J.J., Cortes, T., Martorell, X., Ayguade, E., Labarta, J.: Running OpenMP applications efficiently on an everything-shared SDSM. J. Parallel Distrib. Comput. (May 2006)Google Scholar
- 10.Ferrer, R., Planas, J., Bellens, P., Duran, A., Gonzalez, M., Martorell, X., Badia., R., Ayguade, E., Labarta, J.: Optimizing the Exploitation of Multicore Processors and GPUs with OpenMP and OpenCL. In: Proceedings of the 23rd International Workshop on Languages and Compilers for Parallel Computing (LCPC 2010) (October 2010)Google Scholar
- 11.OpenMP ARB. OpenMP Application Program Interface, v. 3.0 (May 2008)Google Scholar
- 12.Josep, M., Perez, R.M.: Badia, and Jesus Labarta. A dependency-aware task-based programming environment for multi-core architectures. In: IEEE Int. Conference on Cluster Computing, pp. 142–151 (September 2008)Google Scholar
- 14.Rico, A., Duran, A., Cabarcas, F., Ramirez, A., Etsion, Y., Valero, M.: Trace-driven Simulation of Multithreaded Applications. In: Proceedings of the 2011 ISPASS (to appear, 2011)Google Scholar
- 15.Ayguadé, E., Marjanovic, V., Labarta, J., Valero, M.: Effective communication and computation overlap with hybrid mpi/smpss. In: Proceedings of the 15th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP 2010, pp. 337–338. ACM, New York (2010)Google Scholar