Synthesis of ESI Equivalence Class Combinational Circuit

\Permission to make digital/hard copy of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for proot or commercial advantage , the copyright notice, the title of the publication and its date appear, and notice is given that copying is by permission of CBL. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior speciic permission and/or a fee." If you choose to cite this report, please add the following entry to your bibliography database: number = "1997-TR@CBL-07-Harlow", month = "Oct", year = "1997", note = "{Also available at Abstract { Despite more than a decade of experience with the use of standardized benchmark circuits, meaningful comparisons of EDA algorithms remain elusive. In this paper, we introduce an entirely new methodology for characterizing the performance of Binary Decision Diagram (BDD) software. Our method involves the synthesis of large equivalence classes of Entropy{Signature Invariant (ESI) circuits, based on a known reference circuit. We demonstrate that such classes induce controllable distributions of BDD algorithm performance, which provide the foundation for statistically signiicant comparison of diierent algorithms. Introduction The characterization and comparison of Design Automation algorithms has historically been based on the use of widely{ accepted sets of benchmark circuitss1, 2]. The use of standardized test cases creates an impression of generality and lack of bias in comparisons, but in fact, there is little generality in this process. The commonly used benchmark circuits do not represent a continuum of complexity, and in general, the benchmark sets have no invariant properties: each circuit is unique in size, function and architecture. Furthermore, since these circuits are widely available and thoroughly studied, it is likely that many current DA tools are, in eeect, \tuned" to run well on these circuits. The authors do not mean to suggest any deception or dishonesty. The mere use of a small sample of speciic, unrelated circuits as measures of performance inherently induces bias. In most other elds of science, comparisons between theories are routinely made by use of controlled experiments. Acceptance of results by the community is conditioned on demonstration that a theory is repeatable and statistically signiicant; publication is often denied in cases of minor or in-cremental advances. The eld of Electronic Design Automation , in sharp contrast to such rigor, does not have any process analogous to \clinical …