Simulation of Nano-Confined Fluids and Colloids Using GPUs

The nano-confinement of fluids and/or colloids is a particularly interesting problem with applications ranging from lubrication in computer hard drives to gas adsorption in industrial processes.  Understanding the fundamental properties, dynamics, structures, and phase transitions of nano-confined systems is of great importance to the design and engineering of these varied systems and is an area where simulation has been widely applied [1].  In this work we perform MD simulations of nano-confined fluids using the HOOMD-Blue simulation package, optimized to run on graphical processing units (GPUs) [2].   We find single GPU performance equivalent to 64+ commodity processors, allowing for the rapid exploration of a wide range of parameters and the simulation of a large number of statistical measures for each statepoint.  Here, we report the behavior and GPU performance of a variety of different fluid models, ranging from simple isotropic Lennard-Jones spheres to anisotropic fluids to detailed polymer models.

* Funding and computational support provided by the U.S. Civilian Research and Development Foundation, Award Number UKC1-9201-LV-09 and the Keeneland GPU cluster as part of the the National Institute for Computational Sciences, project-ID UT-TNEDU014.

[1] Cummings PT, Docherty H, Iacovella CR, Singh JK, Phase transitions in nanoconfined fluids: The evidence from simulation and theory, AICHE Journal, 56 (4), 2010
[2] Anderson JA, Lorenz, CD, Travesset A, General purpose molecular dynamics simulations fully implemented on graphics processing units, Journal of Computational Physics, 227, (2008)