Fast Boolean Minimizer for Completely Specified Functions
Petr Fišer, Premysl Rucky, Irena Vanova · 2008
We propose a simple and fast two-level minimization algorithm for completely specified functions in this paper. The algorithm is based on processing ternary trees. A ternary tree is proposed as a structure enabling a very compact representation of completely specified Boolean functions. It is efficient especially for functions having many on-set terms. The minimization algorithm is thus most suited for functions described by many on-set terms. Such functions emerge as a result of many algorithms used in logic synthesis process, e.g., multi-level network collapsing, algebraic manipulation with logic functions, etc. When these functions are to be minimized, most of the state- of-the-art minimizers (Espresso) need prohibitively long time to process them, or they are even completely unusable, due to their very high memory consumption. Our algorithm is able to minimize such functions in a reasonable time, though the result quality does not reach the quality of other minimizers. Here our minimizer found its application as a pre-processor that, when run prior to, e.g., Espresso, significantly reduces total minimization time, while fully retaining the result quality.