Systolic architecture for solving NP-hard combinatorial problems of logic design and related areas
Philip Ho, Marek A. Perkowski · 2003
A new approach to solving various NP-hard problems in logic synthesis, logic programming, graph theory, and related areas is presented. A problem to be solved is reduced to solving one or several generic combinatorial problems; this is called generalized propositional formula (GPF) minimization. A special massively parallel computer architecture for GPF minimization is discussed. The architecture is composed of a host computer and a data-flow tree of processors (Boolean product processor, or BPP). Each BPP consists of a product management unit and a sorting and absorbing architecture.>