Indonesia Conference Directory

Corresponding Author
Radhinka Bagaskara

Institutions
a) Master Program in Computational Science, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, Bandung 40132, Indonesia
b) Department of Physics, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, Bandung 40132, Indonesia
*radhinka[at]students.itb.ac.id

Abstract
Molecular dynamics is a method to simulate movements of atoms or particles in a system. The simulation is computationally complex and the performance heavily depends with the computer. As such, a high performance computation system is necessary to achieve the optimal efficiency. One of the solutions is by using parallel computation to speed up the simulation running time. In this research, we use CUDA API developed by Nvidia, to create a parallel molecular dynamics simulation with the help of Nvidias own GPUs. The system used in this simulation is the noble gas particles, i.e. Helium, Neon, Argon, Krypton, Xenon, and Radon. Interactions between particles is modeled by using Lennard-Jones potential and the equations of motion which are integrated with velocity Verlet method. The system is then benchmarked by making comparisons of the execution time between the serial and parallel simulation on CUDA cores. The running time of the parallel simulation is faster than the serial one.

Keywords
Molecular Dynamics, Noble Gas, Parallel Computation, CUDA

Topic
High Performance Computing

Link: https://ifory.id/abstract/DjM3fvKbAJzT