首页> 外国专利> Non-linear, gain-based modeling of circuit delay for an electronic design automation system

Non-linear, gain-based modeling of circuit delay for an electronic design automation system

机译:电子设计自动化系统中基于非线性,基于增益的电路延迟建模

摘要

A non-linear, gain-based modeling of circuit delay within an electronic design automation system. The present invention provides a scalable cell model for use in early logic structuring and mapping for the design of integrated circuits. The scalable cell model includes a four dimensional delay model accepting input slew and gain and providing delay and output slew. By eliminating output loading as a requirement for delay computations, the scalable model of the present invention can effectively be used to provide accurate delay information for early logic synthesis processes, e.g., that precede technology dependent optimizations where the actual load of a cell is unknown. This scalable cell model considers: the impact of transition times on delay; complex gates having different input capacitances for different input pins; the impact of limited discrete cell sizes in the technology library; and design rules, e.g., maximum capacitance and maximum transition associated with gates. A technology library is analyzed and clustering is performed to select a cluster of cells for each cell group of a common functionality. A nominal input slew value is computed for all cells and a scaling factor is computed for each cell of each cluster. From each cluster, a four dimensional gain-based non-linear scalable cell model (look-up table) is generated. A default gain is computed for each scalable cell model and an area model and an input pin capacitance model are generated for each scalable cell model.
机译:电子设计自动化系统中电路延迟的非线性,基于增益的建模。本发明提供了一种可扩展的单元模型,用于集成电路设计的早期逻辑结构和映射。可伸缩单元模型包括一个四维延迟模型,该模型接受输入转换和增益,并提供延迟和输出转换。通过消除作为延迟计算要求的输出负载,本发明的可扩展模型可以有效地用于为早期逻辑合成过程提供准确的延迟信息,例如,在依赖于技术的最优化之前,其中单元的实际负载是未知的。这种可扩展的信元模型考虑:过渡时间对延迟的影响;对于不同的输入引脚,具有不同输入电容的复杂栅极;技术库中有限的离散像元大小的影响;设计规则,例如与门相关的最大电容和最大过渡。分析技术库并执行聚类以为具有公共功能的每个单元组选择一个单元集群。为所有单元计算标称输入转换值,并为每个群集的每个单元计算比例因子。从每个群集中,生成一个基于四维增益的非线性可伸缩单元模型(查找表)。为每个可伸缩单元模型计算默认增益,并为每个可伸缩单元模型生成面积模型和输入引脚电容模型。

著录项

  • 公开/公告号US6543036B1

    专利类型

  • 公开/公告日2003-04-01

    原文格式PDF

  • 申请/专利权人 SYNOPSYS INC.;

    申请/专利号US19990452056

  • 发明设计人 MAHESH IYER;ASHISH KAPOOR;

    申请日1999-11-30

  • 分类号G06F175/00;

  • 国家 US

  • 入库时间 2022-08-22 00:04:37

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号