公开(公告)号：US07284212B2

公开(公告)日：2007-10-16

申请号：US10710509

申请日：2004-07-16

Applicant: Gaurav Kumar Varshney ,  Sreeram Chandrasekar

Inventor： Gaurav Kumar Varshney ,  Sreeram Chandrasekar

IPC: G06F17/50

CPC classification number: G06F17/5081 ,  G06F17/5036 ,  G06F17/5045

Abstract: Reducing the number of computations required to pre-characterize cells in a cell-library. In an embodiment, a worst case vector which propagates most noise on an arc (combination of input pin and output pin) of a cell is determined, and NP characteristics and NIC are generated only for the worst case vector. Noise analysis is then performed using such curves generated from the worst case vector. Since curves corresponding to only the worst case vector may need to be generated, the computational requirements may be reduced. The search ranges in determining the immunity transition points forming the NIC may be reduced, according to some aspects of the present invention. The data corresponding to NIC may be used to generate NP curves, and vice versa to reduce computational requirements further.

Abstract translation: 减少预先表征细胞库中细胞所需的计算次数。 在一个实施例中，确定传播单元的电弧（输入引脚和输出引脚的组合）上的大部分噪声的最坏情况向量，并且仅针对最坏情况矢量生成NP特性和NIC。 然后使用从最坏情况向量生成的这种曲线执行噪声分析。 由于可能需要生成对应于最差情况矢量的曲线，所以可以减少计算要求。 根据本发明的一些方面，可以减少确定形成NIC的免疫转变点的搜索范围。 对应于NIC的数据可以用于产生NP曲线，反之亦然，以进一步减少计算量。