当前位置:首页 > 报告详情

DC25_SLIDES_TRACK07_GeneticAlgorithmDrivenIBISAMI_Auge_V2.pdf

上传人: S** 编号:1240856 2026-05-16 26页 2.23MB

1、Information Classification:GeneralWelcome to ConferenceJanuary 2830,2025Santa Clara Convention Center1ExpoJanuary 2930,2025 Information Classification:GeneralGenetic Algorithm-Driven IBIS-AMI Optimization for Robust 200G SerDes DesignAdrien Auge,(Alphawave Semi)Tripp Worrell,(MathWorks)Walter Katz(M

2、athWorks),Richard Allred(MathWorks)Eric Brock(MathWorks)2Information Classification:GeneralSPEAKERSAdrien AugeSignal Integrity Engineer,Alphawave S|Adrien is a Signal Integrity Engineer at Alphawave Semi and has over 10 years experience related to modeling and simulation of electrical interface.From

3、 parallel and serial interface design for memory and GPU products at SK Hynix and Intel.He also worked at Synopsys on IBIS-AMI modeling for high-speed DSP based SerDes.His current role at Alphawave Semi is to provide accurate SI models to IP customers for their off-chip design needs and enablement,a

4、s well as signal integrity integration support.Tripp WorrellDevelopment Manager,MathW|Tripp is a Development Manager at MathWorks,overseeing SerDes Toolbox,Mixed-Signal Blockset,and Signal Integrity Toolbox.He joined in 2017 after managing EDA software development at SiSoft.Previously,he spent seven

5、 years at Cisco as a Signal Integrity Engineer.Tripp holds a BS in Computer and Electrical Engineering and an MS in Computer Engineering from NC State University.3Information Classification:GeneralIntro&MotivationCurrent modeling&ChallengesOptimization Adaptation limitation&improvementConclusionAgen

6、da4Information Classification:GeneralIBIS-AMI model?5COMSeasimotherIBIS-AMIFullarchitecturalmodelabstractionruntimeInformation Classification:GeneralSerDes Diagram30%of total EQ35%of total EQ35%of total EQtransmitterreceiverInformation Classification:GeneralAdaptation 7TX EQRX EQbackplaneRX Adaptati

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。
1. **会议主题**:2025年1月28-30日圣克拉拉会议中心,聚焦200G SerDes设计中基于遗传算法的IBIS-AMI优化。 2. **核心问题**:200G SerDes行为建模中,均衡器(EQ)优化复杂度高(200G达800万种组合),传统暴力搜索耗时(4.9天)。 3. **解决方案**:遗传算法(GA)显著提升效率,200G优化仅需5.5分钟(900次评估),比暴力搜索快200倍以上。 4. **改进方向**:提出基于波形的新成本函数,解决低损耗通道中非线性失真问题,提升低/中损耗性能。 5. **结论**:GA优化技术可扩展至更复杂SerDes模型,增强IBIS-AMI模型的鲁棒性和适应性。
**GA优化优势?** **IBIS-AMI挑战?** **200G SerDes设计?**
客服
商务合作
小程序
服务号
折叠