AutoSizer: Automatic Sizing of Analog and Mixed-Signal Circuits via Large Language Model (LLM) Agents

2026-02-02

AI Summary

This work addresses the challenges of transistor sizing in analog and mixed-signal (AMS) circuits—namely high-dimensional nonlinearity, stringent constraints, and heavy reliance on expert knowledge—which render traditional EDA methods inefficient and fragile. To overcome these limitations, we propose AutoSizer, the first reflective meta-optimization framework that synergistically integrates large language models (LLMs) with numerical optimization. AutoSizer employs a dual-loop architecture that enables closed-loop automated sizing through circuit-aware reasoning, simulation-feedback-driven dynamic search space refinement, and adaptive optimization scheduling. Our contributions include establishing a novel LLM-agent-driven paradigm for AMS optimization and introducing AMS-SizingBench, the first standardized and open-sourced benchmark for this task. Experimental results demonstrate that AutoSizer substantially outperforms both conventional approaches and existing LLM-based agents across diverse complex circuits, achieving significant improvements in solution quality, convergence speed, and success rate.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Optimization
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search

Abstract

The design of Analog and Mixed-Signal (AMS) integrated circuits remains heavily reliant on expert knowledge, with transistor sizing a major bottleneck due to nonlinear behavior, high-dimensional design spaces, and strict performance constraints. Existing Electronic Design Automation (EDA) methods typically frame sizing as static black-box optimization, resulting in inefficient and less robust solutions. Although Large Language Models (LLMs) exhibit strong reasoning abilities, they are not suited for precise numerical optimization in AMS sizing. To address this gap, we propose AutoSizer, a reflective LLM-driven meta-optimization framework that unifies circuit understanding, adaptive search-space construction, and optimization orchestration in a closed loop. It employs a two-loop optimization framework, with an inner loop for circuit sizing and an outer loop that analyzes optimization dynamics and constraints to iteratively refine the search space from simulation feedback. We further introduce AMS-SizingBench, an open benchmark comprising 24 diverse AMS circuits in SKY130 CMOS technology, designed to evaluate adaptive optimization policies under realistic simulator-based constraints. AutoSizer experimentally achieves higher solution quality, faster convergence, and higher success rate across varying circuit difficulties, outperforming both traditional optimization methods and existing LLM-based agents.

Problem

• Analog and Mixed-Signal circuits
• transistor sizing
• design automation
• optimization bottleneck
• high-dimensional design space

Innovation

• LLM agents
• analog circuit sizing
• meta-optimization
• adaptive search space
• AMS-SizingBench