Accurate Computation of Longest Sensitizable Paths using Answer Set Programming

Abstract:
Precise knowledge of the longest sensitizable paths in a circuit is crucial for various tasks in computer-aided design, including timing analysis, performance optimization, delay testing, and speed binning. As delays in today’s nanoscale technologies are increasingly affected by statistical parameter variations, there is significant interest in obtaining sets of paths that are within a length range. We present an ASP-based method for computing well-defined sets of sensitizable paths within a length range. Unlike previous approaches, the method is accurate and does not rely on a priori relaxations. Experimental results demonstrate the applicability and scalability of our method.