Parallelization of an Advection-Diffusion Problem Arising in Edge Plasma Physics Using Hybrid MPI/OpenMP Programming

Matthieu Kuhn (1) , Guillaume Latu (2) , Nicolas Crouseilles (3, 4) , Stéphane Genaud (1)

1 ICube - Laboratoire des sciences de l'ingénieur, de l'informatique et de l'imagerie
2 IRFM - Institut de Recherche sur la Fusion par confinement Magnétique
3 IPSO - Invariant Preserving SOlvers
4 IRMAR - Institut de Recherche Mathématique de Rennes

This work presents a hybrid MPI/OpenMP parallelization strategy for an advection-diffusion problem, arising in a scientific application simulating tokamak's edge plasma physics. This problem is the hotspot of the system of equations numerically solved by the application. As this part of the code is memory-bandwidth limited, we show the benefit of a parallel approach that increases the aggregated memory bandwidth in using multiple computing nodes. In addition, we designed some algorithms to limit the additional cost, induced by the needed extra inter nodal communications. The proposed solution allows to achieve good scalings on several nodes and to observe 70% of relative efficiency on 512 cores. Also, the hybrid parallelization allows to consider larger domain sizes, unreachable on a single computing node.

Mots clés: Hybrid MPI/OpenMP; Advection-Diffusion; Plasma physics; equation; code

Domaines: Analyse numérique [math.NA]

Matthieu Kuhn, Guillaume Latu, Nicolas Crouseilles, Stéphane Genaud. Parallelization of an Advection-Diffusion Problem Arising in Edge Plasma Physics Using Hybrid MPI/OpenMP Programming. 21st International Conference on Parallel and Distributed Computing (Euro-Par), Vienna University of Technology, Aug 2015, Vienne, Austria. pp.545-557, ⟨10.1007/978-3-662-48096-0_42⟩. ⟨hal-01243542⟩