A Symbolic Simulation Approach
In Resolving Signals' Correlation


Julien Dunoyer, Nizar Abdallah, and Pirouz Bazargan Sabet
Laboratoire MASI/CAO-VLSI, IBP, CNRS-URA818
University Pierre et Marie Curie (PARIS VI)
4, place Jussieu, 75252 Paris Cedex 05 FRANCE



Abstract

To date, the probabilistic approach is in widespread use to analyze the power dissipation of digital circuits. The goal of the present work is twofold. First, to present a method in providing an evaluation package of signals' probabilities and transition densities that can be applied at different specification levels with more or less accuracy. A first order clue at the specification level may help synthesis tools in lowering power dissipation early in the design process. Second, to show how to deal with the well-known correlation problem using a symbolic simulation technique. Only one pass is required to compute a set of independent inputs for each signal in the circuit. At different stages, the algorithms take advantage out of the Binary Decision Diagram (BDD) concept.