Timing-Driven Placement

This chapter reviews two basic sets of netweighting algorithms: static netweighting and dynamic netweighting. It explores two global placement approaches: partitioning and force-directed, a several detailed placement approaches. The chapter examines a mincut-based approach and two analytical partitioning-based approaches and presents several representative and techniques for Timing-driven placement (TDP) and timing-aware placement. Timing convergence metrics measure the extent to which a placement satisfies timing constraints. The most interesting aspect of the TDP is the mechanism to translate timing metrics into actions to drive the core placement engines. Netweighting-based TDP is very simple to implement and less computational intensive. Static netweighting computes the netweights once before TDP. It can be divided into two categories: empirical netweighting and sensitivity-based netweighting. Empirical netweighting assigns netweight based on the critically of the net, which indicates how much the placer should reduce the wirelength on this net.