CROP: Fast and effective congestion refinement of placement

Modern circuits become harder to route with the ever decreasing design features. Previous routability-driven placement techniques are usually tightly coupled with the underlying placers. So usually they cannot be easily integrated into various placement tools. In this paper, we propose a tool called CROP (Congestion Refinement of Placement) for mixed-size placement solutions. CROP is independent of any placer. It takes a legalized placement solution and then relocates the modules to improve routability without significantly disturbing the original placement solution. CROP interleaves a congestion-driven module shifting technique and a congestion-driven detailed placement technique. Basically the shifting technique targets at better allocating the routing resources. Shifting in each direction can be formulated as a linear program (LP) for resizing each G-Cell. Instead of solving the computationally expensive LP, we discover that the LP formulation could be relaxed and solved by a very efficient longest-path computation. Then the congestion-driven detailed placement technique is proposed to better distribute the routing demands. Congestion reduction is realized by weighting the HPWL with congestion coefficient during detailed placement. The experimental results show that CROP is capable of effectively alleviating the congestion for unroutable placement solutions. We apply it to placement solutions generated by four different placers on the ISPD05/06 placement benchmarks. Within a very short runtime, CROP greatly improves the routability and saves execution time for the routing stage after refinement.