Hybrid Parallel Programming Models for AMR Neutron Monte-Carlo Transport
CEA, DAM, DIF, F-91297, Arpajon
This paper deals with High Performance Computing (HPC) applied to neutron transport theory on complex geometries, thanks to both an Adaptive Mesh Refinement (AMR) algorithm and a Monte-Carlo (MC) solver. Several Parallelism models are presented and analyzed in this context, among them shared memory and distributed memory ones such as Domain Replication and Domain Decomposition, together with Hybrid strategies. The study is illustrated by weak and strong scalability tests on complex benchmarks on several thousands of cores thanks to the petaflopic supercomputer Tera100.
Key words: Neutronics / AMR / MPI / OpenMP / Domain Decomposition & Replication / Hybrid Parallelism
© Owned by the authors, published by EDP Sciences, 2014