Constructive AIG optimization considering input weights

This work presents a technology independent synthesis algorithm that works on top of an And-Inverter Graph (AIG) data structure. The goal is to minimize the number of nodes and the depth of the AIG, by controlling these characteristics while building the graph. Moreover, large circuits are usually synthesized in stages (cuts) and, therefore, it is desirable to consider different weights to the input nodes of the AIGs in order to represent signal delay from previous circuit stages. This paper incorporates the employ of input weights to a novel approach for AIG construction based on a new synthesis paradigm called functional composition. The idea is to perform the synthesis by associating simpler AIGs, in a bottom-up approach. The method constructively controls characteristics of final and intermediate graphs, allowing the adoption of secondary criteria other than the number of AIG nodes for optimization.