Generalized symmetric and generalized pseudo-symmetric functions

We introduce generalized symmetric and generalized pseudo-symmetric functions which can be represented as regular two-dimensional linear arrays. It is possible due in part to generalized symmetries, which can be used when functions are represented using exor-based decompositions. We also identify and use local symmetries. Therefore, a large group of non-symmetric functions can be implemented using the same efficient regular locally-connected structures which are used for totally symmetric functions. Because of the regularity, layout parameters such as area, interconnect length and delay can be very accurately estimated during the prelayout synthesis stages, which is critically important for deep-submicron technologies