A new method of identifying critical paths for performance optimization

[[abstract]]It is shown that during performance optimization some false paths may become sensitizable and that the selection of critical paths needs to consider some false paths besides sensitizable paths. These critical paths are called candidate paths. Techniques for extracting candidate paths, and an effective graph model for representing and manipulating the huge number of candidate paths have been developed. The speed and memory utilization of the proposed method are demonstrated using ISCAS benchmark circuits[[fileno]]2030223030011[[department]]資訊工程學