The Algorithm of Control Program Generation for Optimization of LuNA Program Execution
LuNA fragmented programming system is a high-level declarative system of parallel programming. Such systems have the problem of achieving on appropriate program execution performance in comparison with MPI. The reasons are a high degree of parallel program execution non-determinism and execution overhead. The paper presents an algorithm of control program generation for LuNA programs. That is a step towards automatic improvement of LuNA program execution performance. Performance tests presented show effectiveness of the proposed approach.
KeywordsHigh performance computing Fragmented programming technology Fragmented programming system LuNA Parallel program generation
The author would like to thank his supervisor Dr. Victor E. Malyshkin for his professional guidance and Vladislav A. Perepelkin for his constructive suggestions during the development of this research work.
- 5.Bosilca, G., Bouteiller, A., et al.: DAGuE: a generic distributed DAG engine for high performance computing. In Proceedings of IPDPS 2011 Workshops, pp. 1151–1158 (2011)Google Scholar
- 6.Bosilca, G., Bouteiller, A., et al.: Flexible development of dense linear algebra algorithms on massively parallel architectures with DPLASMA. In: Proceedings of IPDPS 2011 Workshops, pp. 1432–1441 (2011)Google Scholar
- 7.Perez, J.M., Badia, R.M., Labarta, J.: A flexible and portable programming model for SMP and multi-cores. Technical report 03/2007, Barcelona Supercomputing Center (2007)Google Scholar
- 8.Kale, L.V., Krishnan, S.: CHARM++: a portable concurrent object oriented system based on C++. In: Proceedings of OOPSLA 1993, pp. 91–108. ACM, New York (1993)Google Scholar
- 9.Huang, C., Laxmikant, V.K.: Charisma: orchestrating migratable parallel objects. In: Proceedings of the 16th International Symposium on High Performance Distributed Computing (HPDC), pp. 75–84 (2007)Google Scholar