C to Cellular Automata and Execution on CPU, GPU and FPGA
Jens Drieseberg and Christian Siemers
Over the last decades Cellular Automata (CA) have become more and more present in solving general-purpose problems, but the main issue is how to map a problem to a Cellular Automata model. Special languages were developed for programming such models, but learning a new programming language or even programming paradigm is very time consuming. Furthermore software developers have to keep in mind the specific structure of Cellular Automata when designing a new algorithm. In this paper we present a method to generate Cellular Automata models from standard C code. The code is transcoded by mapping the complete algorithm written in C to a Cellular Automata model that may be compiled for CPU, GPU and even FPGA without further user interaction.
Keywords: Cellular automata models & algorithms, fine-grained parallel architectures and FPGA, efficient architectures and implementations