High Performance N-Body Simulation and Visualization through CUDA Architecture

Edgar Josafat Martinez Noriega, Tetsu Narumi
National Polytechnic Institute (IPN), Mexico City, Mexico
Bulletin of The University of Electro-Communications, 24-1, pp. 59-64, 2012


   title={High Performance N-Body Simulation and Visualization through CUDA Architecture},

   author={NORIEGA, E.J.M. and NARUMI, T.},



Download Download (PDF)   View View   Source Source   



General purpose computing on graphics processing unit (GPGPU) has become into a new paradigm to program easily massive parallel processors. This hardware architecture is very suitable to encourage N-body problems such as molecular dynamics simulations, due to compute each body on every thread. Visualizations of molecular systems such as ‘claret’ simulator have been developed. However real-time simulation require a lot of computation on the physics and visualizations of atoms. First, we propose an implementation of force calculation and time integration on molecular dynamics NaCl ions through compute unified device architecture (CUDA) on the graphics processor unit (GPU). Second, we decrease the visualization bottlenecks on sphere generation of particles computing each vertex on GPU. On a conventional system (Intel Core 2 Duo 2.6 Ghz) with a GPU (Nvidia Geforce 280 GTX) , we reach 242.4 Gflops during the force-time integration calculation on 2,744 particles and 30.1 frames/sec on sphere rendering (1,481,760 triangles).
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2021 hgpu.org

All rights belong to the respective authors

Contact us: