Behavioral synthesis of digital filters using attribute grammars

Recently, a formal framework to perform behavioral synthesis using attribute grammars has been presented, its main advantages being modularity and declarative notation in the development of design automation environments. From a practical point of view, most modern behavioral synthesizers are best suited for the development of special purpose hardware to implement digital signal processing algorithms. In this paper, the attribute grammar formalism and the corresponding framework are extended to handle the development of special purpose hardware for both FIR and IIR filters. The proposed methodology is capable to construct hardware implementations in various technologies (FPGA, CPLD, ASIC) from behavioral filter specifications (difference equations or discrete convolution), utilizing the VHDL standard hardware description language. Overall, a formal specification for fast and efficient filter implementation is proposed.