Assessing Large Language Models in Generating RTL Design Specifications

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Huang, Hung-Ming, Yang, Yu-Hsin, Chang, Fu-Chieh, Hsu, Yun-Chia, Lin, Yin-Yu, Tsai, Ming-Fang, Yang, Chun-Chih, Wu, Pei-Yuan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912737694580736
author Huang, Hung-Ming
Yang, Yu-Hsin
Chang, Fu-Chieh
Hsu, Yun-Chia
Lin, Yin-Yu
Tsai, Ming-Fang
Yang, Chun-Chih
Wu, Pei-Yuan
author_facet Huang, Hung-Ming
Yang, Yu-Hsin
Chang, Fu-Chieh
Hsu, Yun-Chia
Lin, Yin-Yu
Tsai, Ming-Fang
Yang, Chun-Chih
Wu, Pei-Yuan
contents As IC design grows more complex, automating comprehension and documentation of RTL code has become increasingly important. Engineers currently should manually interpret existing RTL code and write specifications, a slow and error-prone process. Although LLMs have been studied for generating RTL from specifications, automated specification generation remains underexplored, largely due to the lack of reliable evaluation methods. To address this gap, we investigate how prompting strategies affect RTL-to-specification quality and introduce metrics for faithfully evaluating generated specs. We also benchmark open-source and commercial LLMs, providing a foundation for more automated and efficient specification workflows in IC design.
format Preprint
id arxiv_https___arxiv_org_abs_2512_00045
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Assessing Large Language Models in Generating RTL Design Specifications
Huang, Hung-Ming
Yang, Yu-Hsin
Chang, Fu-Chieh
Hsu, Yun-Chia
Lin, Yin-Yu
Tsai, Ming-Fang
Yang, Chun-Chih
Wu, Pei-Yuan
Hardware Architecture
Artificial Intelligence
As IC design grows more complex, automating comprehension and documentation of RTL code has become increasingly important. Engineers currently should manually interpret existing RTL code and write specifications, a slow and error-prone process. Although LLMs have been studied for generating RTL from specifications, automated specification generation remains underexplored, largely due to the lack of reliable evaluation methods. To address this gap, we investigate how prompting strategies affect RTL-to-specification quality and introduce metrics for faithfully evaluating generated specs. We also benchmark open-source and commercial LLMs, providing a foundation for more automated and efficient specification workflows in IC design.
title Assessing Large Language Models in Generating RTL Design Specifications
topic Hardware Architecture
Artificial Intelligence
url https://arxiv.org/abs/2512.00045