arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.00386cs.AR
收藏

AssertLLM: 通过多语言模型生成和评估硬件验证断言

AssertLLM: Generating and Evaluating Hardware Verification Assertions from Design Specifications via Multi-LLMs

Hong Kong University of Science and Technology(香港科技大学);Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州))

Wenji Fang, Mengming Li, Min Li, Zhiyuan Yan, Shang Liu, Hongce Zhang, Zhiyao Xie

AI总结:

AssertLLM通过多语言模型自动从设计规范生成准确的硬件验证断言,有效提升验证效率和准确性。

摘要:

基于断言的验证(ABV)是确保设计电路符合其架构规范的关键方法,这些规范通常以自然语言描述。这一过程通常需要验证工程师进行人工解释,将这些规范转换为功能验证断言。现有方法在从自然语言规范生成断言方面仅限于工程师提取的句子,阻碍了其实际应用。在本工作中,我们提出了AssertLLM,一个自动断言生成框架,能够处理完整的规范文件。AssertLLM将复杂任务分解为三个阶段,结合三个定制的大型语言模型(LLMs)来提取结构规范、映射信号定义并生成断言。我们对AssertLLM在完整设计上的评估,涵盖23个I/O信号,显示生成的89%断言在语法和功能上都是准确的。

英文摘要:

Assertion-based verification (ABV) is a critical method for ensuring design circuits comply with their architectural specifications, which are typically described in natural language. This process often requires human interpretation by verification engineers to convert these specifications into functional verification assertions. Existing methods for generating assertions from natural language specifications are limited to sentences extracted by engineers, discouraging its practical application. In this work, we present AssertLLM, an automatic assertion generation framework that processes complete specification files. AssertLLM breaks down the complex task into three phases, incorporating three customized Large Language Models (LLMs) for extracting structural specifications, mapping signal definitions, and generating assertions. Our evaluation of AssertLLM on a full design, encompassing 23 I/O signals, demonstrates that 89\% of the generated assertions are both syntactically and functionally accurate.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑