一种新的基于晶体管级的电路划分算法

随着VLSI电路规模的不断增加,为实现电路并行仿真所做的电路划分算法的质量显得日益重要。鉴于现有算法未能同时保证均衡的分块间规模和最少的互联信号数目,该文提出了一种新的基于晶体管级的电路划分算法。该算法首先通过一个聚合过程对电路网表进行分割,得到一个比较好的初始分割;然后通过平衡分块间规模差异和进一步优化分块间互连线的数目,最终得到理想的电路划分结果。应用该电路划分算法对工业界的实际电路网表进行测试,结果表明:相比于目前普遍使用的COPART算法,该算法在分块间规模的均衡性方面平均改善了25%,在分块间的互联信号数目方面平均减少了18%。     

应用于65nm CMOS技术节点漏电流分析的精确快速的查表模型

Novel physical models for leakage current analysis in 65 nm technology are proposed. Taking into consideration the process variations and emerging effects in nano-scaled technology, the presented models are capable of accurately estimating the subthreshold leakage current and junction tunneling leakage current in 65 nm technology. Based on the physical models, new table look-up models are developed and first applied to leakage current analysis in pursuit of higher simulation speed. Simulation results show that the novel physical models are in excellent agreement with the data measured from the foundry in the 65 nm process, and the proposed table look-up models can provide great computational efficiency by using suitable interpolation techniques. Compared with the traditional physical-based models, the table look-up models can achieve 2.5X speedup on average on a variety of industry circuits.     

Accurate and fast table look-up models for leakage current analysis in 65 nm CMOS technology

Novel physical models for leakage current analysis in 65 nm technology are proposed. Taking into con-sideration the process variations and emerging effects in nano-scaled technology, the presented models are capable of accurately estimating the subthreshold leakage current and junction tunneling leakage current in 65 nm technol-ogy. Based on the physical models, new table look-up models are developed and first applied to leakage current analysis in pursuit of higher simulation speed. Simulation results show that the novel physical models are in ex-cellent agreement with the data measured from the foundry in the 65 nm process, and the proposed table look-up models can provide great computational efficiency by using suitable interpolation techniques. Compared with the traditional physical-based models, the table look-up models can achieve 2.5X speedup on average on a variety of industry circuits.