The Improved VLSI Floorplanalgorithm And Its Realization Based On Single Sequence

The design process of VLSI circuit is a complex and time-consuming work. The VLSI CAD tool is introduced to save the design time, human and material resources, so it is very important to the development of integrated circuit technology.This thesis introduces the purpose and significance of VLSI floorplanning, and the general process of the VLSI physical design; describes the Single Sequence representation for the floorplan, and the correspondence relationship between the floorplan and the related Single Sequence. Based on this description, this thesis describes the traditional floorplanning algorithms based on the Single Sequence. Furthermore, describes the algorithm principle, process and simulation results emphatically. Firstly the phenomena of multiple sequences corresponding to the same floorplan is studied which is the traditional floorplanning algorithms based on the Single Sequence briefly. After that, the concept of the authenticity of the Single Sequence is introduced. Based on the conception, the entire list of Single Sequence is discussed which analyze the root cause of the problem and the optimization results. Finally, based on the previous analysis and comparison results, the existing algorithm is improved which can solve the problem that multiple sequence corresponding to the same floorplan. The simulation results also show that the efficiency of improved algorithm has been greatly improved, so this method should have a very good value. Then, this thesis introduces a conception of flexibility and weight of the module, and a new mathod for module placement is introduced which follow the strategy that the maximum weight module first. Based on the maximum weight first floorplanning algorithm, a improved algorithm is introduced which can dynamically adjust the module selection probability in accordance with the module weight and the current annealing temperature.Also,The simulation results show that the improved algorithm optimise the convergence speed and improves the efficiency of algorithm comparing with other algorithm.