AMS-IO-Bench and AMS-IO-Agent: Benchmarking and Structured Reasoning for Analog and Mixed-Signal Integrated Circuit Input/Output Design

2025-12-25

AI Summary

A significant gap exists between natural-language design intent and industrial-grade implementation in analog/mixed-signal (AMS) IC I/O design. Method: This paper introduces AMS-IO-Agent—the first domain-specific large language model (LLM) agent for AMS chip I/O ring design—integrating structured intent modeling with a reusable constraint knowledge base, employing dual intermediate representations (JSON/Python), a structured knowledge graph, and deep integration with DRC/LVS verification and industrial EDA toolchains. Contribution/Results: We release AMS-IO-Bench, the first automated benchmark suite for AMS I/O design. For the first time, an LLM-generated I/O ring has undergone silicon validation and tape-out at 28 nm. Experiments demonstrate >70% DRC+LVS pass rate and reduce design turnaround time from hours to minutes, establishing a new human–machine collaborative paradigm for silicon-proven I/O design.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent Communication

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

In this paper, we propose AMS-IO-Agent, a domain-specialized LLM-based agent for structure-aware input/output (I/O) subsystem generation in analog and mixed-signal (AMS) integrated circuits (ICs). The central contribution of this work is a framework that connects natural language design intent with industrial-level AMS IC design deliverables. AMS-IO-Agent integrates two key capabilities: (1) a structured domain knowledge base that captures reusable constraints and design conventions; (2) design intent structuring, which converts ambiguous user intent into verifiable logic steps using JSON and Python as intermediate formats. We further introduce AMS-IO-Bench, a benchmark for wirebond-packaged AMS I/O ring automation. On this benchmark, AMS-IO-Agent achieves over 70% DRC+LVS pass rate and reduces design turnaround time from hours to minutes, outperforming the baseline LLM. Furthermore, an agent-generated I/O ring was fabricated and validated in a 28 nm CMOS tape-out, demonstrating the practical effectiveness of the approach in real AMS IC design flows. To our knowledge, this is the first reported human-agent collaborative AMS IC design in which an LLM-based agent completes a nontrivial subtask with outputs directly used in silicon.

Problem

• Automating analog/mixed-signal I/O subsystem generation from natural language design intent
• Converting ambiguous user requirements into verifiable structured design steps
• Reducing AMS IC I/O ring design time from hours to minutes

Innovation

• LLM-based agent for AMS IC I/O generation
• Structured domain knowledge base with reusable constraints
• Design intent structuring using JSON and Python formats