Memory Affinity for Hierarchical Shared Memory Multiprocessors

Abstract : Currently, parallel platforms based on large scale hierarchical shared memory multiprocessors with Non-Uniform Memory Access (NUMA) are becoming a trend in scientific High Performance Computing (HPC). Due to their memory access constraints, these platforms require a very careful data distribution. Many solutions were proposed to resolve this issue. However, most of these solutions did not include optimizations for numerical scientific data (array data structures) and portability issues. Besides, these solutions provide a restrict set of memory policies to deal with data placement. In this paper, we describe an user-level interface named Memory Affinity interface (MAi), which allows memory affinity control on Linux based cache-coherent NUMA (ccNUMA) platforms. Its main goals are, fine data control, flexibility and portability. The performance of MAi is evaluated on three ccNUMA platforms using numerical scientific HPC applications, the NAS Parallel Benchmarks and a Geophysics application. The results show important gains (up to 31%) when compared to Linux default solution.
Liste complète des métadonnées
Contributor : Arnaud Legrand <>
Submitted on : Friday, February 15, 2013 - 1:46:21 PM
Last modification on : Thursday, October 11, 2018 - 8:48:02 AM

Links full text




Christiane Pousa Ribeiro, Marcio Bastos Castro, Luiz Gustavo Fernandes, Alexandre Carissimi, Jean-François Mehaut. Memory Affinity for Hierarchical Shared Memory Multiprocessors. Proceedings of IEEE International Symposium on Computer Architectures and High Performance Computing (SBAC-PAD), 2009, Sao Paulo, Brazil. pp.59-66, ⟨10.1109/SBAC-PAD.2009.16⟩. ⟨hal-00788914⟩



Record views