Improving Memory Affinity of Geophysics Applications on NUMA platforms Using Minas
Christiane Pousa Ribeiro (University of Grenoble - INRIA - Mescal Research Team)
Márcio Bastos Castro (University of Grenoble - INRIA - Mescal Research Team)
Jean-François Méhaut (University of Grenoble - INRIA - Mescal Research Team)
Alexandre Carissimi (Universidade Federal do Rio Grande do Sul)
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.
Parallel and Distributed Computing, Performance Analysis