A true congestion prediction method based on router's intelligence

This paper presents a true probabilistic congestion prediction method based on router's intelligence to be used in placement stage of physical design flow. The bend distributions for IBM-PL ACE benchmarks were extracted after detailed routing and used to guide the probabilistic method. Furthermore, router's behavior for congestion minimization was modeled and used in our prediction method. The experimental results show that for IBM-PLACE benchmarks, our estimation method predicted the congestion more accurately than a recent method by about 19%.