Scheduling Dynamic OpenMP Applications over Multicore Architectures Export

@incollection{citeulike:3255760, abstract = {Approaching the theoretical performance of hierarchical multicore machines requires a very careful distribution of threads and data among the underlying non-uniform architecture in order to minimize cache misses and NUMA penalties. While it is acknowledged that OpenMP can enhance the quality of thread scheduling on such architectures in a portable way, by transmitting precious information about the affinities between threads and data to the underlying runtime system, most OpenMP runtime systems are actually unable to efficiently support highly irregular, massively parallel applications on NUMA machines. In this paper, we present a thread scheduling policy suited to the execution of OpenMP programs featuring irregular and massive nested parallelism over hierarchical architectures. Our policy enforces a distribution of threads that maximizes the proximity of threads belonging to the same parallel region, and uses a NUMA-aware work stealing strategy when load balancing is needed. It has been developed as a plug-in to the forestGOMP OpenMP platform\ [TBG+07]. We demonstrate the efficiency of our approach with a highly irregular recursive OpenMP program resulting from the generic parallelization of a surface reconstruction application. We achieve a speedup of 14 on a 16-core machine with no application-level optimization.}, author = {Broquedis, Fran\c{c}ois and Diakhat\'{e}, Fran\c{c}ois and Thibault, Samuel and Aumage, Olivier and Namyst, Raymond and Wacrenier, Pierre-Andr\'{e}}, citeulike-article-id = {3255760}, citeulike-linkout-0 = {http://dx.doi.org/10.1007/978-3-540-79561-2_15}, doi = {10.1007/978-3-540-79561-2_15}, journal = {OpenMP in a New Era of Parallelism}, keywords = {multi-thread, scheduling}, pages = {170--180}, posted-at = {2008-09-15 03:39:14}, priority = {2}, title = {Scheduling Dynamic OpenMP Applications over Multicore Architectures}, url = {http://dx.doi.org/10.1007/978-3-540-79561-2_15}, year = {2008} }

TY - CHAP ID - citeulike:3255760 L3 - citeulike-article-id:3255760 N2 - Approaching the theoretical performance of hierarchical multicore machines requires a very careful distribution of threads and data among the underlying non-uniform architecture in order to minimize cache misses and NUMA penalties. While it is acknowledged that OpenMP can enhance the quality of thread scheduling on such architectures in a portable way, by transmitting precious information about the affinities between threads and data to the underlying runtime system, most OpenMP runtime systems are actually unable to efficiently support highly irregular, massively parallel applications on NUMA machines. In this paper, we present a thread scheduling policy suited to the execution of OpenMP programs featuring irregular and massive nested parallelism over hierarchical architectures. Our policy enforces a distribution of threads that maximizes the proximity of threads belonging to the same parallel region, and uses a NUMA-aware work stealing strategy when load balancing is needed. It has been developed as a plug-in to the forestGOMP OpenMP platform [TBG+07]. We demonstrate the efficiency of our approach with a highly irregular recursive OpenMP program resulting from the generic parallelization of a surface reconstruction application. We achieve a speedup of 14 on a 16-core machine with no application-level optimization. JF - OpenMP in a New Era of Parallelism EP - 180 TI - Scheduling Dynamic OpenMP Applications over Multicore Architectures SP - 170 KW - multi-thread KW - scheduling AU - Broquedis, François AU - Diakhaté, François AU - Thibault, Samuel AU - Aumage, Olivier AU - Namyst, Raymond AU - Wacrenier, Pierre-André PY - 2008/// UR - http://dx.doi.org/10.1007/978-3-540-79561-2_15 ER -