🤖 AI Summary
This paper systematically investigates the adaptation pathways and application boundaries of large language models (LLMs) in electronic design automation (EDA). Addressing critical challenges—including significant semantic gaps between LLMs and EDA tasks, difficulties in domain-knowledge integration, and insufficient end-to-end coverage—we propose the first comprehensive classification framework for deep LLM–EDA integration. Our methodology introduces a customized paradigm encompassing architectural evolution, scaling laws, task-specific modeling, knowledge-augmented reasoning, and domain-aware prompt engineering. Through empirical evaluation across frontend synthesis, physical design, and verification tasks, we derive an extensible LLM application taxonomy, precisely delineating capability boundaries and requirement-alignment mechanisms. We identify six fundamental technical challenges and articulate concrete, practice-oriented research directions. This work establishes both theoretical foundations and actionable technical roadmaps for deploying LLMs in EDA.
📝 Abstract
Within the rapidly evolving domain of Electronic Design Automation (EDA), Large Language Models (LLMs) have emerged as transformative technologies, offering unprecedented capabilities for optimizing and automating various aspects of electronic design. This survey provides a comprehensive exploration of LLM applications in EDA, focusing on advancements in model architectures, the implications of varying model sizes, and innovative customization techniques that enable tailored analytical insights. By examining the intersection of LLM capabilities and EDA requirements, the paper highlights the significant impact these models have on extracting nuanced understandings from complex datasets. Furthermore, it addresses the challenges and opportunities in integrating LLMs into EDA workflows, paving the way for future research and application in this dynamic field. Through this detailed analysis, the survey aims to offer valuable insights to professionals in the EDA industry, AI researchers, and anyone interested in the convergence of advanced AI technologies and electronic design.