作者: Jhih-Rong Gao , Pei-Ci Wu , Ting-Chi Wang
关键词: Identification (information) 、 Routing (electronic design automation) 、 Iterative method 、 Process (computing) 、 Parallel computing 、 Integrated circuit design 、 Computer science 、 Computer network 、 Benchmark (computing) 、 Router 、 Bottleneck
摘要: In this paper, we present a new global router, NTHU-Route, for modern designs. NTHU-Route is based on iterative rip-ups and reroutes, several techniques are proposed to enhance our router. These include (1) history cost function which helps distribute overflow during (2) an adaptive multi-source multi-sink maze routing method improve the wirelength of routing, (3) congested region identification specify order nets be ripped up rerouted, (4) refinement process further reduce when reroutes reach bottleneck. Compared with two state-of-the-art works ISPD98 benchmarks, outperforms them in both wirelength. For much larger designs from ISPD07 benchmark suite, solution quality better than or comparable best results reported contest.