Cache Aware Dynamics Data Layout for Efficient Shared Memory Parallelisation of EUROPLEXUS

Marwa Sridi (1, 2) , Bruno Raffin (1) , Vincent Faucher (2)

1 DATAMOVE - Data Aware Large Scale Computing
2 CEA - Commissariat à l'énergie atomique et aux énergies alternatives

Parallelizing industrial simulation codes like the EUROPLEXUS software dedicated to the analysis of fast transient phenomena, is challenging. In this paper we focus on the efficient parallelization on a multi-core shared memory node. We propose to have each thread gather the data it needs for processing a given iteration range, before to actually advance the computation by one time step on this range. This lazy cache aware layout construction enables to keep the original data structure and leads to very localised code modifications. We show that this approach can improve the execution time by up to 40% when the task size is set to have the data fit in the L2 cache.

Mots clés: EUROPLEXUS; Shared Memory; Cache-aware Data Layout; Parallel Programming

Domaines: Calcul parallèle, distribué et partagé [cs.DC]

Marwa Sridi, Bruno Raffin, Vincent Faucher. Cache Aware Dynamics Data Layout for Efficient Shared Memory Parallelisation of EUROPLEXUS. International Conference on Computational Science (ICCS), Jun 2016, San Diego, United States. pp.1083 - 1092, ⟨10.1016/j.procs.2016.05.413⟩. ⟨hal-01420005⟩