RTLocating: Intent-aware RTL Localization for Hardware Design Iteration

2026-02-27

AI Summary

This work addresses the limitation of existing LLM-based approaches to hardware design, which typically support only one-shot RTL generation and struggle to accommodate iterative, natural language-driven updates required in industrial chip development. To enable intent-aware, localized RTL modifications, we propose RTLocating—a novel framework that formalizes the ΔSpec-to-RTL localization task for the first time. RTLocating integrates three complementary views: semantic meaning, local structural context, and global dependency patterns, through a multi-view adaptive fusion architecture featuring a GLIDE (Global Interaction and Local Dependency Encoding) module and a dynamic routing mechanism. We also introduce EvoRTL-Bench, the first industrial-scale benchmark aligning natural language change intents with corresponding RTL code segments. On this benchmark, RTLocating achieves state-of-the-art performance with 0.568 MRR and 15.08% R@1, outperforming the strongest baseline by 22.9% and 67.0%, respectively.

Technology Category

• Machine Learning: Hardware-aware ML
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Local Search

Application Category

• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Industrial chip development is inherently iterative, favoring localized, intent-driven updates over rewriting RTL from scratch. Yet most LLM-Aided Hardware Design (LAD) work focuses on one-shot synthesis, leaving this workflow underexplored. To bridge this gap, we for the first time formalize $Δ$Spec-to-RTL localization, a multi-positive problem mapping natural language change requests ($Δ$Spec) to the affected Register Transfer Level (RTL) syntactic blocks. We propose RTLocating, an intent-aware RTL localization framework, featuring a dynamic router that adaptively fuses complementary views from a textual semantic encoder, a local structural encoder, and a global interaction and dependency encoder (GLIDE). To enable scalable supervision, we introduce EvoRTL-Bench, the first industrial-scale benchmark for intent-code alignment derived from OpenTitan's Git history, comprising 1,905 validated requests and 13,583 $Δ$Spec-RTL block pairs. On EvoRTL-Bench, RTLocating achieves 0.568 MRR and 15.08% R@1, outperforming the strongest baseline by +22.9% and +67.0%, respectively, establishing a new state-of-the-art for intent-driven localization in evolving hardware designs.

Problem

• RTL localization
• hardware design iteration
• intent-aware
• ΔSpec-to-RTL
• change request

Innovation

• RTL localization
• intent-aware design
• dynamic routing
• hardware design iteration
• EvoRTL-Bench