HDLdebugger: Streamlining HDL debugging with Large Language Models

Xufeng Yao,Haoyang Li,Tsz Ho Chan,Wenyi Xiao,Mingxuan Yuan,Yu Huang,Lei Chen,Bei Yu

2024-03-18

Abstract:In the domain of chip design, Hardware Description Languages (HDLs) play a pivotal role. However, due to the complex syntax of HDLs and the limited availability of online resources, debugging HDL codes remains a difficult and time-intensive task, even for seasoned engineers. Consequently, there is a pressing need to develop automated HDL code debugging models, which can alleviate the burden on hardware engineers. Despite the strong capabilities of Large Language Models (LLMs) in generating, completing, and debugging software code, their utilization in the specialized field of HDL debugging has been limited and, to date, has not yielded satisfactory results. In this paper, we propose an LLM-assisted HDL debugging framework, namely HDLdebugger, which consists of HDL debugging data generation via a reverse engineering approach, a search engine for retrieval-augmented generation, and a retrieval-augmented LLM fine-tuning approach. Through the integration of these components, HDLdebugger can automate and streamline HDL debugging for chip design. Our comprehensive experiments, conducted on an HDL code dataset sourced from Huawei, reveal that HDLdebugger outperforms 13 cutting-edge LLM baselines, displaying exceptional effectiveness in HDL code debugging.

Hardware Architecture,Artificial Intelligence,Computational Engineering, Finance, and Science,Machine Learning,Software Engineering

What problem does this paper attempt to address?

### Problems the paper attempts to solve In the field of chip design, Hardware Description Languages (HDLs) play a crucial role. However, due to the complex grammar of HDL and limited online resources, debugging HDL code is a difficult and time - consuming task even for experienced engineers. Therefore, there is an urgent need to develop automated HDL code debugging models to relieve the burden on hardware engineers. Although large - language models (LLMs) have shown strong capabilities in generating, completing, and debugging software code, in the specialized field of HDL debugging, the application of LLMs is still limited and has not yet achieved satisfactory results. This paper proposes an HDL debugging framework based on large - language models - HDLdebugger, which aims to automate and simplify the HDL debugging process through the following three components: 1. **Data Generation**: Generate HDL debugging data through reverse - engineering methods. 2. **Search Engine**: A search engine for Retrieval - Augmented Generation (RAG). 3. **Retrieval - enhanced LLM Fine - tuning**: Fine - tune the LLM through retrieval - enhanced methods. Through the integration of these components, HDLdebugger can automate and simplify HDL debugging in chip design. Experimental results show that HDLdebugger significantly outperforms 13 state - of - the - art LLM baseline models on the HDL code dataset provided by Huawei, demonstrating its excellent performance in HDL code debugging.

HDLdebugger: Streamlining HDL debugging with Large Language Models

MetaHDL:inference and parameter tracing oriented domain-specific language for hardware description

LLM-Aided Efficient Hardware Design Automation

LDB: A Large Language Model Debugger via Verifying Runtime Execution Step-by-step

LLM4SecHW: Leveraging Domain Specific Large Language Model for Hardware Debugging

Classification-Based Automatic HDL Code Generation Using LLMs

From English to ASIC: Hardware Implementation with Large Language Model

HiVeGen -- Hierarchical LLM-based Verilog Generation for Scalable Chip Design

VHDL-Eval: A Framework for Evaluating Large Language Models in VHDL Code Generation

Enhancing the Code Debugging Ability of LLMs via Communicative Agent Based Data Refinement

LLM4EDA: Emerging Progress in Large Language Models for Electronic Design Automation

MEIC: Re-thinking RTL Debug Automation using LLMs

Automated C/C++ Program Repair for High-Level Synthesis via Large Language Models

HLSPilot: LLM-based High-Level Synthesis

Self-HWDebug: Automation of LLM Self-Instructing for Hardware Security Verification

Optimizing High-Level Synthesis Designs with Retrieval-Augmented Large Language Models

VerilogReader: LLM-Aided Hardware Test Generation

A Unified Debugging Approach via LLM-Based Multi-Agent Synergy

Towards LLM-Powered Verilog RTL Assistant: Self-Verification and Self-Correction

Digital ASIC Design with Ongoing LLMs: Strategies and Prospects

LLHD: A Multi-level Intermediate Representation for Hardware Description Languages