Abstract

Cellular automata (CA), an inherently parallel computing paradigm, have been successfully applied to a range of computational problems to model the behavior of complex environmental systems. Field Programmable Gate Arrays (FPGA) are an ideal platform to investigate the hardware (HW) realization of CA models. FPGA based CA accelerators are not new to the scientific community, however mapping CA models to the HW logic remains a major challenge. In the present work we report on computationally intensive CA-based algorithms and their implementation on FPGA. We highlight intrinsic similarities between CA and FPGA and discuss possible constraints and trade-offs towards appropriate mapping of CA-based models and better exploitation of the favorable HW characteristics. We also refer to various FPGA realizations of our CA environmental systems that deliver significant performance evaluation compared to the corresponding software solutions.