Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Zhishuai Zhang, Xintian Li, Shilong Liu, Aodong Zhang, Lu Jie, Nan Sun

TL;DR

AMS-IO-Agent tackles the automation gap in AMS IC I/O design by grounding LLM reasoning with a domain-specific knowledge base and a structured design-intent graph to produce production-ready EDA scripts. It introduces AMS-IO-Bench as a practical benchmark and demonstrates strong performance, achieving over $70\%$ DRC+LVS signoff in bench tests and validating a wirebond-packaged I/O ring in a real 28-nm tape-out. The results show substantial reductions in design turnaround time from hours to minutes and robust integration into industrial flows, surpassing a baseline LLM. The work establishes a practical, generalizable framework for human-agent collaboration in AMS IC design and paves the way for broader application of domain-specialized LLMs in hardware design.

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.

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

TL;DR

AMS-IO-Agent tackles the automation gap in AMS IC I/O design by grounding LLM reasoning with a domain-specific knowledge base and a structured design-intent graph to produce production-ready EDA scripts. It introduces AMS-IO-Bench as a practical benchmark and demonstrates strong performance, achieving over DRC+LVS signoff in bench tests and validating a wirebond-packaged I/O ring in a real 28-nm tape-out. The results show substantial reductions in design turnaround time from hours to minutes and robust integration into industrial flows, surpassing a baseline LLM. The work establishes a practical, generalizable framework for human-agent collaboration in AMS IC design and paves the way for broader application of domain-specialized LLMs in hardware design.

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.
Paper Structure (21 sections, 5 figures, 5 tables)

This paper contains 21 sections, 5 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. LLM-based Approaches for IC Design Automation
  4. Benchmark for AMS IC agents
  5. Automation in I/O Ring Assembly
  6. Method
  7. Task Definition
  8. Agent Overview
  9. Design Intent Structuring
  10. Intent Graph Adaptor
  11. Domain-specific Knowledge Base
  12. Benchmark
  13. Experiments
  14. Setup
  15. Evaluation Metrics
  16. ...and 6 more sections

Figures (5)

  • Figure 1: Comparison of conventional and agent-driven I/O ring design in a real AMS IC design workflow.
  • Figure 2: Comparison of baseline direct code generation and the proposed structured design intent.
  • Figure 3: Overview of the proposed AMS-IO-Agent architecture, illustrating the workflow and the component roles.
  • Figure 4: Illustration of the intent graph and examples of its nodes.
  • Figure 5: Application of the proposed agent system to a real AMS tape-out I/O ring design.