符号化矩量与敏感度方法用于抗串扰噪声布线-电路与系统专业论文.docx

第 第 II 页 SYMBOLIC MOMENT AND MOMENT SENSITIVITY COMPUTATION METHOD WITH APPLICATION TO CROSSTALK AVOIDANCE IN ROUTING ABSTRACT As the VLSI feature size scaled down, the crosstalk and coupling effects between Interconnects become more and more significant and the chip is facing a serious signal integrity problem. At the same time, die size and power dissipation also directly affects the performance and reliability of the chip. Wire sizing has been proved been a good method on physical optimization. Due to high computation complexity, most traditional wire-sizing algorithm use simple mathematical model to estimate the signal delay and the effect of coupling. This paper proposes a symbolic moment sensitivity computation method, which can be used to compute moment sensitivities versus all RC elements in the circuit by traverse the symbolic representation only once. Symbolic Moment and Moment Sensitivity evaluation method has the advantage of high efficiency and repeatability, which was applied to solve a more precise physical optimization model by the Author. The optimization model uses D2M as the metric of signal delay, account the worst case that aggressor can affect the victim when coupling is considered. The experiment result shows that such a model has higher fidelity when compared to traditional wire-sizing models. KEY WORDS: Lagrangian Relaxation、Symbolic、Moment Sensitivity、 Wire-Sizing、crosstalk、 第 第 PAGE IV 页 目 录 第 1 章 绪论 1 1.1 引言 1 1.2 互连线模型及研究 2 1.2.1 互连线的传输线模型 2 1.2.2 互连线的集总 RC 模型 3 1.2.3 互连线的 RLC 模型 5 1.2.4 工艺尺寸缩小对互连线参数的影响 6 1.3 互连线上的信号完整性 7 1.3.1 互连线的串扰估计算法 7 1.3.2 互连线的信号延迟 13 1.3.3 互连线的延迟估计 16 1.3.4 符号化矩计算 17 1.4 以互连线为中心的设计流程 19 1.5 VLSI 物理设计中的优化 21 1.5.1 时序优化 21 1.5.2 信号完整性优化 22 1.5.3 面积优化 22 1.6 VLSI 物理优化算法 23 1.7 本文的主要研究内容及安排 25 第 2 章 符号化矩敏感度计算 27 2.1 矩敏感度介绍 27 2.2 符号化加权矩敏感度计算 28 2.2.1 矩敏感度计算的递归表达式推导 29 2.2.2 关于电阻元件的矩敏感度计算的符号化表示 30 2.2.3 关于电容元件的矩敏感度计算的符号化表示 34 2.3 符号化矩敏感度计算的复杂度分析 36 2.4 符号化敏感度计算的正