AVERAGE POWER ANALYSIS OF SEQUENTIAL-CIRCUITS USING AN AUTOREGRESSIVEMODEL

We present a new statistical technique for average power estimation in sequential circuits. Because of the feedback loops, power dissipation s of sequential circuits in consecutive clock cycles are temporally co rrelated. The existence of data correlation makes it unsuitable to app ly conventional techniques to average power inference, because the sam ple variance is no longer a maximum likelihood estimator. The converge nce criterion derived from the biased variance estimation will be over ly optimistic, causing power simulation to stop prematurely at a lower -than-specified estimation accuracy. To overcome this problem, we prop ose a systematic approach for modeling the power dissipation behavior of sequential circuits as an autoregressive random process. An accurat e process variance can be obtained by the model parameters, which enab les the derivation of a robust confidence interval of the average powe r. The interval is checked for convergence against a user-specified ac curacy criterion. An iterative procedure is developed to invoke these steps repeatedly until the convergence specification is met. For a set of benchmark sequential circuits, this technique yields high accuracy and efficiency.