Application of Indexed Partition Calculus in Logic Synthesis of Boolean Functions for FPGAs

Functional decomposition of Boolean functions specified by cubes proved to be very efficient. Most popular decomposition methods are based on blanket calculus. However computation complexity of blanket manipulations strongly depends on number of function's variables, which prevents them from being used for large functions of many input and output variables. In this paper a new concept of indexed partition is proposed and basic operations on indexed partitions are defined. Application of this concept to logic synthesis based on functional decomposition is also discussed. The experimental results show that algorithms based on new concept are able to deliver good quality solutions even for large functions and does it many times faster than the algorithms based on blanket calculus.