Implementation of FHP model in Nvidia CUDA technology.

My Master's Thesis in pdf file to download (in polish).

Poster presentation in pdf file to download.


The aim of the study is to examine the possibility of accelerating FHP algorithms used for hydrodynamics simulations using the newest Nvidia CUDA parallel programming technology.

The first chapter presents an introduction to the cellular automata and the concept of using it as a model of discretization of time and space. It leads to the most advanced FHP III model and its implementation.

The second and the third chapters describe the technology used in the study. The second chapter is an introduction to the Nvidia CUDA programming model and the third one explains how to use it for FHP implementation. The GPU over CPU speedup is presented and discussed.

The last chapter verifies the correctness of the FHP algorithm implemented for this study. The results were compared to the theoretical solution of the Navier-Stokes equations and experimental results obtained by the group from Max Planck Institute for Marine Microbiology (A. Khalili).

Example simulations:

Example usage of my program is shown below. It is written in C with CUDA,OpenGL,Glut. On GTX260 it generates about 150fps with ~10^6 node lattice.

Flow past a cylinder, with von karman vertex.

Secondary flow in corners of square: