Parallelization of the Symplectic Massive Body Algorithm (SyMBA) N-body Code

Abstract

Direct N-body simulations of a large number of particles, especially in the study of planetesimal dynamics and planet formation, have been computationally challenging even with modern machines. This work presents the combination of fully parallelized N 2/2 interactions and the incorporation of the GENGA code’s close encounter pair grouping strategy to enable MIMD parallelization of the Symplectic Massive Body Algorithm (SyMBA) with OpenMP on multi-core CPUs in shared-memory environment. SyMBAp (SyMBA parallelized) preserves the symplectic nature of SyMBA and shows good scalability, with a speedup of 30.8 times with 56 cores in a simulation with 5000 fully interactive particles.