IncreRTL: Traceability-Guided Incremental RTL Generation under Requirement Evolution

arXiv cs.AI / 3/30/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper presents IncreRTL, an LLM-driven method for generating RTL code incrementally when design requirements evolve, rather than regenerating entire modules.
IncreRTL builds requirement-to-code traceability links to identify which RTL segments are affected by requirement changes and only regenerates those parts.
The approach aims to reduce structural drift that can occur with prior static LLM RTL generation methods and to improve both update accuracy and engineering consistency.
Evaluation on the newly created EvoRTL-Bench shows IncreRTL delivers better regeneration consistency and efficiency versus existing strategies.
The work is positioned as a step toward practical deployment of LLM-based RTL generation in real engineering workflows with iterative requirement changes.

Abstract

Large language models (LLMs) have shown promise in generating RTL code from natural-language descriptions, but existing methods remain static and struggle to adapt to evolving design requirements, potentially causing structural drift and costly full regeneration. We propose IncreRTL, a LLM-driven framework for incremental RTL generation under requirement evolution. By constructing requirement-code traceability links to locate and regenerate affected code segments, IncreRTL achieves accurate and consistent updates. Evaluated on our newly constructed EvoRTL-Bench, IncreRTL demonstrates notable improvements in regeneration consistency and efficiency, advancing LLM-based RTL generation toward practical engineering deployment.