Cardiac tissue simulation using graphics hardware

As video cards become faster and more programmable, physical simulations implemented on graphics processors become possible. This paper examines different programmable stages of the nVidia graphics processor (GPLI), and their use to simulate electrical activation of cells in a tissue sample using a cellular automaton model. Comparable tissue simulation programs were written to run on both the CPU and the GPU. A simple activation rule set was used to provide a base for speed evaluation of the graphics processor vs. a standard personal computer. The results show that a speed increase of 1.6/spl times/ is achieved for simulations of 2048 /spl times/ 2048 cells. CPU's offer a low cost high speed platform to conduct tissue simulations.