当前位置:首页 > 报告详情

具有网络排序优化的成对间距约束封装布线.pdf

上传人: 芦苇 编号:651781 2025-05-01 38页 1.32MB

1、Paired-Spacing-Constrained Package Routing with Net Ordering OptimizationPresenter:Ying-Jie JiangAdvisor:Shao-Yun Fang12OutlineIntroductionAlgorithm FlowNet Ordering DeterminationTile and Network Graph ConstructionMCMF-based Global Routing DerivationExperimental ResultsConclusion3Introduction4Introd

2、uctionPaired-spacing constraint is one of the most important constraints in advanced package technologyA net may require different spacing depending on the nets it is adjacent toThere can be at most 2 spacing rules for netsIt is important to establish different spacing rules for different bump netsE

3、nsure signal integrity Minimize crosstalk5Example of Pair SpacingIf we apply post-refinement or a pessimistic spacing ruleResult in suboptimal solution:Normal Net Bump:Critical Spacing:Normal Spacing:Critical Net Bump:Ball:Spacing Violation(a)(b)(c)(d)6Design RulesNon-crossing ConstraintTwo nets can

4、not be routed across each otherRouting Direction ConstraintRouted in four kinds of orientationVertical,horizontal,45,and 135Routing Angle ConstraintReversed 45 turn is not allowedPaired Spacing ConstraintFor normal(non-critical)nets,keep at least a distance from other netsFor critical nets,keep at l

5、east a distance from other nets7Problem FormulationGivenA chip set,a normal bump set,a critical bump set,and a ball grid array(BGA)Wire width Spacing rules for normal bumps and critical bumpsOutputConnect all bumps to ballsMinimized wirelengthWithout any design rule violationFor better readability a

6、nd more concise explanationsOnly consider two spacing constraints in our workCritical spacing rule,normal spacing rule.The proposed algorithm can be trivially extended to general problem instancesArbitrary spacing value can be set between each pair of adjacent nets.8Algorithm Flow9Algorithm Overview

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文提出了一种针对先进封装技术中配对间隙约束的布线算法。该算法通过预处理确定布线顺序,利用动态规划算法优化布线,并在全局布线中采用最大流最小割方法。实验结果显示,该算法在所有案例中均实现了100%的布线可行性,并且与现有技术相比,在布线可行性上提高了13%。此外,该算法还考虑了布线方向、角度和网络图构建,通过添加虚拟点优化了布局,并在详细布线阶段采用了A*搜索算法。该研究展示了在芯片集、正常凸起集、关键凸起集和球格数组中,如何有效处理配对间隙约束,确保信号完整性和最小化交叉干扰。
"如何优化芯片封装中的配对间隙约束?" "基于动态规划的凸点重新排序算法有何优势?" "实验结果证明,新的全局路由算法有何突破?"
客服
商务合作
小程序
服务号
折叠