This repository curates research papers applying Artificial Intelligence (AI) and Machine Learning (ML) to Electronic Design Automation (EDA). It targets engineers, researchers, and power users seeking to leverage AI/ML for automating, optimizing, and accelerating complex chip design workflows, offering a comprehensive overview of the state-of-the-art across various EDA stages.

How It Works

The project compiles research demonstrating the application of diverse ML techniques, including deep learning (GNNs, CNNs), reinforcement learning, and active learning, to address challenges in EDA. These methods are employed for tasks such as design space exploration, performance and resource estimation, placement and routing optimization, and test/verification automation, aiming to improve efficiency and accuracy over traditional EDA flows.

Quick Start & Requirements

This repository is a curated list of research papers and does not provide a software package with installation instructions or a direct quick-start guide. It serves as a reference for academic and industry research in AI for EDA.

Highlighted Details

• Comprehensive coverage spans High-Level Synthesis (HLS) for design space exploration and performance prediction, Logic Synthesis for optimization, and Physical Design stages like Floorplanning, Placement, and Routing for predicting congestion and optimizing layouts.
• Explores AI/ML applications in Testing and Verification, including LLM agents for DRC script synthesis and hybrid methods for analog/mixed-signal circuit verification.
• Features research on accelerating EDA tasks with deep learning engines, such as GPU-accelerated placement tools.
• Includes papers on applying ML to analog circuit design, device sizing, and topology selection.

Maintenance & Community

The collection was compiled by students from Tsinghua University's "Computer-Aided Design of Digital Circuits and Systems" course (2020 Spring). Suggestions can be provided via GitHub issues or by emailing nicsefc@gmail.com.

Licensing & Compatibility

Licensing information for the compiled research papers or any associated code is not specified in the provided README. Compatibility for commercial use or closed-source linking is not addressed.

Limitations & Caveats

This resource is a bibliography of research papers, not a deployable software framework. While it points to numerous AI/ML applications in EDA, it does not directly provide tools, codebases (except for one paper mentioning a code link), or pre-trained models for immediate adoption. Users must consult the individual papers for implementation details and potential limitations.