Abstract
An OpenCL-driven, gpu-based molecular dynamics simulation is described. Using the simulation, statistical thermodynamic properties of a non-ideal gas system are extracted and analyzed in an attempt to extract the second virial coefficient from a system containing 32³ particles. Simulation results agree with theoretical predictions for stability, Maxwell speed and velocity distributions, temperature, and pressure in equilibrium. The relationships between initial conditions and equilibrium conditions are analyzed. The interactions of non-ideal gases have significant impact on these relationships. We also studied the second virial coefficient as a measure of deviation of a non-ideal gas from an ideal gas. The results extracted from simulations show sensitive dependence on the sampling density and temperature ranges. Compared with theoretical predictions, qualitative agreement can be obtained for certain cases, but quantitative agreement can be challenging to obtain numerically.