Efficient decomposition of assigned sequential machines and Boolean functions for PLD implementations

Abstract

Programmable devices impose hard constraints on the size of circuits that can be implemented with them or with their blocks. Therefore, decompositional implementation of large functions is not a choice but a real necessity in this case. The objective of the new decomposition method presented is to implement a complex Boolean function or an assigned sequential machine with a minimal number of constrained building blocks and minimal connections between the blocks. The method aims at solving problems with hard constraints in multiple dimensions. It differs substantially from the methods for solving similar simpler one dimensional problems. The method is very general and, after some small changes, it can be applied to many other partitioning problems. The experimental results show that the method is very efficient.<>