A Robust Routing Guide Generation Approach for Mixed-Size Designs

Routing is a very complex and time-consuming stage in the physical design flow of modern circuits. After routing, the Design Rule Check (DRC) is performed, and the found DRC violations must be eliminated afterward. Therefore, how to guide a router to produce a solution with fewer or no DRC violations (and hence better quality) is desired. This paper presents an approach to automatically generate a set of routing guides from a placement that enables a commercial router to produce a routing solution of better quality. Each routing guide changes the track utilization percentage in a particular area and hence changes the default routing behavior. Our approach adopts machine learning and heuristic techniques to (1) fast predict DRC violation regions from a placement and a set of routing guides and (2) effectively employ the DRC violation region predictor to generate a set of routing guides that assists the router in producing as less DRC violations as possible. Encouraging experimental results are shown to support our approach.