Functional vector generation for combinational circuits based on data path coverage metric and mixed integer linear programming

Abstract

In this paper, a functional vector generation method to maximize the data path coverage of a combinational circuit is introduced. We present a new gate model based on sensitization requirements for transition propagation, and introduce a new methodology to obtain functional vectors of maximum coverage based on Mixed Integer Linear Programming (MILP). Performance comparison and results based on a large set of MCNC'91 [1] benchmark circuits are given. Experimental results show significant speedups over a greedy SAT method.