Effective bounding techniques for solving unate and binate covering problems

Covering problems arise in many areas of electronic design automation such as logic minimization and technology mapping. An exact solution can critically impact both size and performance of the devices being designed. This paper introduces eclipse, a branch-and-bound solver that can solve many covering problems orders of magnitude faster than existing solvers. When used in place of the default covering engine of a well-known logic minimizer, eclipse makes it possible to find, in less than six minutes, true minima for three benchmark problems that have eluded exact solutions for more than a decade.