Power consumption is rapidly becoming an area

No abstract

The state assignment problem for nite state machines has been explored extensively by the logic synthesis and asynchronous design communities. In this paper, we introduce the concepts of base switching and relative switching. We present a fast intelligent exchange based re-encoding algorithm for reducing average switching. The reasons for the speed of this iterative algorithm are given along with an analysis of its time complexity. Experimental results on sequential circuits from the MCNC benchmark set show that switching activity can be reduced by 75%(max) and 35-47%(average). Power savings of 50%(max) and 10-18%(average) were obtained with negligible increase in area. Based on the experimental results we draw conclusions about the best choice of a starting point for the re-encoding algorithm.

No abstract