摘要
Crosstalk has become one of the most critical concerns in very deep sub-micron era. This paper deals with the problem of crosstalk mitigation at both methodological and algorithmic levels. Noting that intermediate operations between global routing and detailed routing are very effective in crosstalk estimation and reduction, the authors propose to incorporate several intermediate steps that are separated in traditional design flow into an integrated routing resource assignment stage, so that the operations could easily cooperate to fully exert their power on crosstalk reduction. An efficient priority-based heuristic algorithm is developed, which works slice by slice. Crosstalk avoidance, and ,nany other aspects that are critical in routing practice including congestion, vias, layer preference, etc., are taken into account. A track reservation strategy is adopted in the algorithm framework to compensate the undesired effects caused by sequential routing. Experimental results on a series of ISPD98 and industrial benchmarks show that the proposed approach is able to reduce capacitive crosstalk by about 70% on average without compromising completion ratio compared with a previously reported graph based algorithm, demonstrating the advantages of the approach.
Crosstalk has become one of the most critical concerns in very deep sub-micron era. This paper deals with the problem of crosstalk mitigation at both methodological and algorithmic levels. Noting that intermediate operations between global routing and detailed routing are very effective in crosstalk estimation and reduction, the authors propose to incorporate several intermediate steps that are separated in traditional design flow into an integrated routing resource assignment stage, so that the operations could easily cooperate to fully exert their power on crosstalk reduction. An efficient priority-based heuristic algorithm is developed, which works slice by slice. Crosstalk avoidance, and ,nany other aspects that are critical in routing practice including congestion, vias, layer preference, etc., are taken into account. A track reservation strategy is adopted in the algorithm framework to compensate the undesired effects caused by sequential routing. Experimental results on a series of ISPD98 and industrial benchmarks show that the proposed approach is able to reduce capacitive crosstalk by about 70% on average without compromising completion ratio compared with a previously reported graph based algorithm, demonstrating the advantages of the approach.
基金
This work is supported by the National Hi-Tech Research ＆ Development 863 Program of China under Grant No. 2004AA1Z14600 and the National Natural Science Foundation of China （NSFC） under Grant No, 60476014.
