一种新的基于单元扩大的拥挤度驱动的布局算法

描述了一种新的基于单元扩大的拥挤度驱动的布局算法 .这个方法用概率估计模型和星型模型来评价线网的走线 .使用全局优化和划分交替的算法来进行总体布局 .提出了单元的虚拟面积的概念 ,单元的虚拟面积不仅体现了单元的面积 ,而且指出了对布线资源的需求 .单元的虚拟面积可以由单元的扩大策略来得到 .把单元的虚拟面积用到划分过程中 ,从而减小拥挤度 .并且使用了单元移动的策略来进一步减小走线的拥挤 .用来自美国公司的一些例子测试了这个算法 ,结果显示布局的结果在可布性方面有了很大的提高

A New Congestion-Driven Placement Algorithm Based on Cell Inflation

A new congestion driven placement Algorithm is described based on the cell inflation. In this approach, the methods of probability estimation and star model are used to evaluate the routing of nets. Global placement can be done by using the algorithm of global optimization and slicing partitioning. The denotation of virtual area of cell is given to indicate not only the area of cell but also the routing demand.The virtual area of a cell is got by using the strategy of cell inflation, with which in the slicing partitioning, the routing congestion is eliminated. Further reduction in congestion is achieved by cell moving. The algorithm has been tested on a set of sample circuits from American companies, with great improvement in routablity having been obtained.