This special session explores transformative advancements in Electronic Design Automation (EDA) methodologies specifically tailored for analog/mixed-signal (AMS) integrated circuits. As modern AMS designs confront escalating challenges from nanometer-scale effects, heterogeneous integration, and stringent power-performance requirements, conventional EDA approaches often prove inadequate. The session will highlight four critical dimensions of innovation: (1) machine learning-enhanced simulation and modeling techniques that accelerate design convergence, (2) novel approaches for handling parasitics and variability in deep-submicron nodes, (3) co-design methodologies bridging analog front-ends with digital back-ends, and (4) signoff-quality verification for AMS circuits in hyperscale SoCs.
本专题论坛探讨针对模拟/混合信号集成电路的电子设计自动化(EDA)方法的突破性进展。随着现代AMS设计面临纳米级效应、异构集成和严格的功耗性能要求等日益严峻的挑战,传统EDA方法往往显得力不从心。本次论坛将重点展示四个关键创新方向:(1)加速设计收敛的机器学习增强仿真与建模技术,(2)处理深亚微米节点寄生效应和变异性的新方法,(3)连接模拟前端与数字后端的协同设计方法学,以及(4)超大规模SoC中AMS电路的签核质量验证技术。