Improving Memory Affinity of Geophysics Applications on NUMA platforms Using Minas

Abstract : On numerical scientific High Performance Computing (HPC), Non-Uniform Memory Access (NUMA) platforms are now commonplace. On such platforms, the memory affinity management remains an important concern in order to overcome the memory wall problem. Prior solutions have presented some drawbacks such as machine dependency and a limited set of memory policies. This paper introduces Minas, a framework which provides either explicit or automatic memory affinity management with architecture abstraction for ccNUMAs. We evaluate our solution on two ccNUMA platforms using two geophysics parallel applications. The results show some performance improvements in comparison with other solutions available for Linux.
Type de document :
Communication dans un congrès
9th International Meeting High Performance Computing for Computational Science (VECPAR), 2010, Berkeley, United States. Springer, 6449, pp.279-292, 2011, 〈10.1007/978-3-642-19328-6_27〉
Liste complète des métadonnées

https://hal.inria.fr/hal-00788872
Contributeur : Arnaud Legrand <>
Soumis le : vendredi 15 février 2013 - 13:09:53
Dernière modification le : mercredi 11 avril 2018 - 01:54:24

Identifiants

Collections

Citation

Christiane Pousa Ribeiro, Marcio Bastos Castro, Jean-François Mehaut, Alexandre Carissimi. Improving Memory Affinity of Geophysics Applications on NUMA platforms Using Minas. 9th International Meeting High Performance Computing for Computational Science (VECPAR), 2010, Berkeley, United States. Springer, 6449, pp.279-292, 2011, 〈10.1007/978-3-642-19328-6_27〉. 〈hal-00788872〉

Partager

Métriques

Consultations de la notice

146