caa a22 4500 445851910 CHVBK 20180317145416.0 cr unu---uuuuu 170323e20110401xx s 000 0 eng 10.1007/s11227-009-0340-3 doi (NATIONALLICENCE)springer-10.1007/s11227-009-0340-3 Region-based parallelization of irregular reductions onexplicitly managed memory hierarchies [Elektronische Daten] [Seonggun Kim, Hwansoo Han, Kwang-Moo Choe] Multicore architectures are evolving with the promise of extreme performance for the classes of applications that require high performance and large bandwidth of memory. Irregular reduction is one of important computation patterns for many complex scientific applications, and it typically requires high performance and large bandwidth of memory. In this article, we propose region-based parallelization techniques for irregular reductions on multicore architectures with explicitly managed memory hierarchies. Managing memory hierarchy in software requires a lot of programming efforts and tends to be error-prone. The difficulties are even worse for applications with irregular data access patterns. To relieve the burden of memory management from programmers, we develop abstractions, particularly targeted to irregular reduction, for structuring parallel tasks, mapping the parallel tasks to processing units and scheduling data transfers between the memory hierarchies. Our framework employs iteration reordering based on regions of data along with dynamic scheduling of parallel tasks. We experimentally evaluate the effectiveness of our techniques for irregular reduction kernels on the Cell processor embedded in a Sony PlayStation3. Experimental results show the speedups of 8 to 14 on the six available SPEs. Springer Science+Business Media, LLC, 2009 Explicitly managed memory hierarchies nationallicence Irregular reduction nationallicence Loop parallelization nationallicence Compilation techniques nationallicence Kim Seonggun Department of Computer Science, KAIST, 305-701, Daejeon, Republic of Korea aut Han Hwansoo Department of Computer Engineering, Sungkyunkwan University, 440-746, Suwon, Republic of Korea aut Choe Kwang-Moo Department of Computer Science, KAIST, 305-701, Daejeon, Republic of Korea aut The Journal of Supercomputing Springer US; http://www.springer-ny.com 56/1(2011-04-01), 25-55 0920-8542 56:1<25 2011 56 11227 https://doi.org/10.1007/s11227-009-0340-3 text/html Onlinezugriff via DOI 1 research-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/s11227-009-0340-3 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Kim Seonggun Department of Computer Science, KAIST, 305-701, Daejeon, Republic of Korea aut NATIONALLICENCE 700 1- Han Hwansoo Department of Computer Engineering, Sungkyunkwan University, 440-746, Suwon, Republic of Korea aut NATIONALLICENCE 700 1- Choe Kwang-Moo Department of Computer Science, KAIST, 305-701, Daejeon, Republic of Korea aut NATIONALLICENCE 773 0- The Journal of Supercomputing Springer US; http://www.springer-ny.com 56/1(2011-04-01), 25-55 0920-8542 56:1<25 2011 56 11227 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer