arXiv cs.AI· Patrik Drazic, Benaoumeur Senouci, Boualem Benatallah·· 13 小时前AI 评分31
VHDLBench：面向 VHDL 芯片设计的大语言模型基准测试与微调
VHDLBench: Benchmarking and Fine-Tuning Large Language Models for VHDL-Based Chip Design
AI 导读
研究团队构建了 VHDLBench，包含 18,500+ 条语法校验样本的 VHDL 数据集，并用 QLoRA 对五个 7B 参数 LLM 进行微调。微调模型在常见设计任务上语法正确率最高达 92%、功能准确率达 73%，其中 DeepSeek-Coder 提升最显著。生成的设计已在 Zynq SoC 上完成 FPGA 原型验证，证明其具备实际可运行性。
正文 · 原文
Abstract:The rapid evolution of digital circuit complexity demands innovative approaches to hardware design automation. While Large Language Models (LLMs) have influenced software development, their application to Hardware Description Language (HDL) generation remains largely unexplored, particularly for VHDL. This paper presents a comprehensive study for AI-assisted digital circuit design, addressing gaps in current Electronic Design Automation workflows. The work introduces a VHDL dataset of 18,500+ syntax-verified examples from open-source repositories and converted Verilog designs. Five 7B-parameter LLMs were fine-tuned using parameter-efficient QLoRA technique, enabling deployment in resource-constrained environments. To evaluate model efficacy, a rigorous three-level pipeline was established: (1) syntax and synthesis validation, (2) functional simulation with testbenches, and (3) FPGA prototyping on Zynq SoCs. Results show fine-tuned models achieve up to 92% syntax correctness and 73% functional accuracy on common design tasks, with DeepSeek-Coder demonstrating the highest performance gains. Generated designs were successfully deployed on real hardware, confirming not only syntactic but operational viability. This work demonstrates that LLMs, when tailored with domain-specific data and lightweight fine-tuning can contribute meaningfully to the future of generative EDA workflows.
| Comments: | This paper is a long version of a paper accepted for the 4th International Conference on Foundation and Large Language Models, 2026 |
| Subjects: | Hardware Architecture (cs.AR); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.28676 [cs.AR] |
| (or arXiv:2610.28676v1 [cs.AR] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28676 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Benaoumeur Senouci [view email]
[v1]
Mon, 19 Oct 2026 12:23:01 UTC (1,488 KB)
来源：arXiv cs.AI · arxiv.org