Аннотация:An approach to use general-purpose graphics processing units (GPGPU) using nVidia CUDA technology1 to increase the performance of cellular automata is presented in this paper. Shown that the calculations on the GPU faster (more than one order of magnitude) in comparison with CPU calculations and a well-functioning C-code.
Cellular automata with Margolus’ neighborhood are used for the demonstration of general steps in CUDA simulations.