Implementing the Global Cellular Automata on CUDA