Understanding and Mitigating Errors of LLM-Generated RTL Code

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Jiazheng, Liu, Cheng, Cheng, Long, Li, Xiaowei, Li, Huawei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914297002590208
author Zhang, Jiazheng
Liu, Cheng
Cheng, Long
Li, Xiaowei
Li, Huawei
author_facet Zhang, Jiazheng
Liu, Cheng
Cheng, Long
Li, Xiaowei
Li, Huawei
contents Despite limited success in large language model (LLM)-based register-transfer-level (RTL) code generation, the root causes of errors remain poorly understood. To address this, we conduct a comprehensive error analysis, finding that most failures arise not from deficient reasoning, but from a lack of RTL programming knowledge, insufficient circuit understanding, ambiguous specifications, or misinterpreted multimodal inputs. Leveraging in-context learning, we propose targeted correction techniques: a retrieval-augmented generation (RAG) knowledge base to supply domain expertise; design description rules with rule-checking to clarify inputs; external tools to convert multimodal data into LLM-compatible formats; and an iterative simulation-debugging loop for remaining errors. Integrating these into an LLM-based framework yields significant improvement, achieving 98.1% accuracy on the VerilogEval benchmark with DeepSeek-v3.2-Speciale, demonstrating the effectiveness of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2508_05266
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Understanding and Mitigating Errors of LLM-Generated RTL Code
Zhang, Jiazheng
Liu, Cheng
Cheng, Long
Li, Xiaowei
Li, Huawei
Hardware Architecture
Computation and Language
Machine Learning
Despite limited success in large language model (LLM)-based register-transfer-level (RTL) code generation, the root causes of errors remain poorly understood. To address this, we conduct a comprehensive error analysis, finding that most failures arise not from deficient reasoning, but from a lack of RTL programming knowledge, insufficient circuit understanding, ambiguous specifications, or misinterpreted multimodal inputs. Leveraging in-context learning, we propose targeted correction techniques: a retrieval-augmented generation (RAG) knowledge base to supply domain expertise; design description rules with rule-checking to clarify inputs; external tools to convert multimodal data into LLM-compatible formats; and an iterative simulation-debugging loop for remaining errors. Integrating these into an LLM-based framework yields significant improvement, achieving 98.1% accuracy on the VerilogEval benchmark with DeepSeek-v3.2-Speciale, demonstrating the effectiveness of our approach.
title Understanding and Mitigating Errors of LLM-Generated RTL Code
topic Hardware Architecture
Computation and Language
Machine Learning
url https://arxiv.org/abs/2508.05266