This paper presents initial results on the problem of allocation of the available VLSI chip's area among various functional components such as I/O pads, memory cells, and internal wiring. First, a general lower bound for any chip computing a transitive function is derived; this bound is tight for certain functions. The arguments used in the various derivations are later used to specify which of the components are critical depending on the relative sizes of the chip and the number of variables of the function to be computed. The general lower bound is powerful enough that many of the previously proved lower bounds (which could account only for some of the functional requirements) are obtained as explicit special cases of the new result.
ASJC Scopus subject areas
- Computer Science(all)