1992•IEEE Transactions on Computer-Aided Design of Integrated Circuits and SystemsRequires access
PLADE: a two-stage PLA decomposition
Maciej Ciesielski, Shipin Yang
Open publisher page 51 citations
Abstract
An efficient method for PLA decomposition, in which a single two-level Boolean function (PLA) is decomposed into two stages of cascaded PLAs such that the total area of all PLAs is smaller than that of the original PLA, is presented. The first stage may contain an arbitrary number of PLAs (generalized decoders), and the second stage contains a single PLA. Primary inputs are partitioned into disjoint sets of input variables and represented as multiple-valued variables so as to minimize the total PLA area. Two efficient algorithms for assignment of input variables to individual decoders are presented, one based on an integer programming technique and the other on graph partitioning. The constrained input encoding of the second-stage PLA is performed by an efficient procedure based on graph coloring and partitioning theory.>