Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SPICEPilot: Navigating SPICE Code Generation and Simulation with AI Guidance

Deepak Vungarala, Sakila Alam, Arnob Ghosh, Shaahin Angizi

TL;DR

This paper presents SPICEPilot—a novel Python-based dataset generated using PySpice, along with its accompanying framework that automates the creation of SPICE simulation scripts, introduces standardized benchmarking metrics to evaluate LLM’s ability for circuit generation, and outlines a roadmap for integrating LLMs into the hardware design process.

Abstract

Large Language Models (LLMs) have shown great potential in automating code generation; however, their ability to generate accurate circuit-level SPICE code remains limited due to a lack of hardware-specific knowledge. In this paper, we analyze and identify the typical limitations of existing LLMs in SPICE code generation. To address these limitations, we present SPICEPilot a novel Python-based dataset generated using PySpice, along with its accompanying framework. This marks a significant step forward in automating SPICE code generation across various circuit configurations. Our framework automates the creation of SPICE simulation scripts, introduces standardized benchmarking metrics to evaluate LLM's ability for circuit generation, and outlines a roadmap for integrating LLMs into the hardware design process. SPICEPilot is open-sourced under the permissive MIT license at https://github.com/ACADLab/SPICEPilot.git.

SPICEPilot: Navigating SPICE Code Generation and Simulation with AI Guidance

TL;DR

This paper presents SPICEPilot—a novel Python-based dataset generated using PySpice, along with its accompanying framework that automates the creation of SPICE simulation scripts, introduces standardized benchmarking metrics to evaluate LLM’s ability for circuit generation, and outlines a roadmap for integrating LLMs into the hardware design process.

Abstract

Large Language Models (LLMs) have shown great potential in automating code generation; however, their ability to generate accurate circuit-level SPICE code remains limited due to a lack of hardware-specific knowledge. In this paper, we analyze and identify the typical limitations of existing LLMs in SPICE code generation. To address these limitations, we present SPICEPilot a novel Python-based dataset generated using PySpice, along with its accompanying framework. This marks a significant step forward in automating SPICE code generation across various circuit configurations. Our framework automates the creation of SPICE simulation scripts, introduces standardized benchmarking metrics to evaluate LLM's ability for circuit generation, and outlines a roadmap for integrating LLMs into the hardware design process. SPICEPilot is open-sourced under the permissive MIT license at https://github.com/ACADLab/SPICEPilot.git.

Paper Structure

This paper contains 12 sections, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background
  3. How tall does LLM stand in SPICE Code generation?
  4. Misconception of Gate Width and Length
  5. Inability to Perform Circuit Analysis
  6. Incorrect Input Signal Assignment and Device Parameter Configuration
  7. SPICEPilot
  8. Framework
  9. Dataset Generation and Benchmarking
  10. Experimental Analysis
  11. Discussions and Future Works
  12. Conclusion

Figures (2)

  • Figure 1: Illustration of the errors generated by LLM for hardware in SPICE. (a) Incorrect W:L ratio, (b) Inability to perform circuit analysis, (c) incorrect input signals.
  • Figure 2: SPICEPilot Framework.