NSF Young Investigator: Performance-Driven VLSI Designs

高性能超大规模集成电路系统设计中的三个研究问题是：(1)如何将以电气参数为特征的系统性能功能与超大规模集成电路物理设计的几何参数联系起来；(2)如何基于给定的技术和计算能力对性能驱动的VLSI物理设计进行建模；(3)如何描述建模问题的基本计算方面，并开发有效的算法来解决它们。设计了用于互连的分布式RLC电路模型。它考虑了电路的非单调响应、信号线之间的耦合效应和低能耗。提出了一种利用自适应小波配置法求解时变麦克斯韦方程组的有效计算方法。这些算法和方法被包含在一个CAD系统中。

Three research issues in high-performance VLSI system design are being addressed: (1) how to relate the system performance function, characterized by electrical parameters, to the geometrical parameters of the VLSI physical design; (2) how to model performance driven VLSI physical designs based on given technology and computational capability; and (3) how to characterize the fundamental computational aspects of modeled problems and develop effective algorithms for solving them. A distributed RLC circuit model for interconnects is being designed. It considers: non-monotone circuit response, coupling effect among the signal lines, and low energy consumption. Efficient computation methods that solve time-varying Maxwell equations using the adaptive wavelet collocation method (AWCM) are being devised. The algorithms and methods are being included in a CAD system.