Advancements in Electronic Design Automation via Large Language Models

The field of electronic design automation (EDA) is experiencing a significant shift with the integration of large language models (LLMs) in various aspects of the design process. Researchers are exploring the potential of LLMs to automate and optimize tasks such as optimization algorithm design, antenna modeling, and high-level synthesis. The use of LLMs is enabling the discovery of new optimization algorithms, improving the efficiency of antenna modeling, and enhancing the accuracy of high-level synthesis predictions. Furthermore, LLMs are being used to develop new frameworks for automated design space exploration, routing-informed placement, and code lifting. These advancements are paving the way for more efficient, scalable, and intelligent EDA workflows. Notable papers in this area include: Evolution of Optimization Algorithms for Global Placement via Large Language Models, which presents an automated framework for evolving optimization algorithms using LLMs. LEAM: A Prompt-only Large Language Model-enabled Antenna Modeling Method, which demonstrates the effectiveness of LLMs in generating accurate antenna models. Intelligent4DSE: Optimizing High-Level Synthesis Design Space Exploration with Graph Neural Networks and Large Language Models, which proposes a framework that integrates a graph neural network with a large language model-enhanced evolutionary algorithm to optimize high-level synthesis design space exploration.

Sources

Evolution of Optimization Algorithms for Global Placement via Large Language Models

LEAM: A Prompt-only Large Language Model-enabled Antenna Modeling Method

Automated Routing-Informed Placement for Large-Scale Photonic Integrated Circuits

Fitness Landscape of Large Language Model-Assisted Automated Algorithm Search

Intelligent4DSE: Optimizing High-Level Synthesis Design Space Exploration with Graph Neural Networks and Large Language Models

Guided Tensor Lifting

BLADE: Benchmark suite for LLM-driven Automated Design and Evolution of iterative optimisation heuristics

LIFT: LLM-Based Pragma Insertion for HLS via GNN Supervised Fine-Tuning

Built with on top of