Dr.~RTL: Autonomous Agentic RTL Optimization through Tool-Grounded Self-Improvement

arXiv cs.AI / 4/17/2026

📰 NewsDeveloper Stack & InfrastructureIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces Dr.~RTL, an agentic framework aimed at automatic RTL (register-transfer level) timing optimization to improve performance, power, and area (PPA), addressing gaps in prior work’s unrealistic evaluations and limited optimization methods.
Dr.~RTL is evaluated in a more realistic setup using larger, more challenging real-world RTL designs and an industrial EDA workflow, rather than relying on manually degraded small designs and weaker open-source tools.
The system runs closed-loop optimization with a multi-agent pipeline for critical-path analysis, parallel RTL rewriting, and tool-based evaluation to iteratively improve timing.
It also proposes group-relative skill learning that distills optimization experience from comparisons of parallel rewrites into an interpretable, reusable skill library, which currently contains 47 pattern–strategy entries and can evolve over time.
In experiments on 20 real-world RTL designs, Dr.~RTL reports average WNS/TNS improvements of 21%/17% along with a 6% area reduction compared with the industry-leading commercial synthesis tool.

Abstract

Recent advances in large language models (LLMs) have sparked growing interest in automatic RTL optimization for better performance, power, and area (PPA). However, existing methods are still far from realistic RTL optimization. Their evaluation settings are often unrealistic: they are tested on manually degraded, small-scale RTL designs and rely on weak open-source tools. Their optimization methods are also limited, relying on coarse design-level feedback and simple pre-defined rewriting rules. To address these limitations, we present Dr. RTL, an agentic framework for RTL timing optimization in a realistic evaluation environment, with continual self-improvement through reusable optimization skills. We establish a realistic evaluation setting with more challenging RTL designs and an industrial EDA workflow. Within this setting, Dr. RTL performs closed-loop optimization through a multi-agent framework for critical-path analysis, parallel RTL rewriting, and tool-based evaluation. We further introduce group-relative skill learning, which compares parallel RTL rewrites and distills the optimization experience into an interpretable skill library. Currently, this library contains 47 pattern--strategy entries for cross-design reuse to improve PPA and accelerate convergence, and it can continue evolving over time. Evaluated on 20 real-world RTL designs, Dr. RTL achieves average WNS/TNS improvements of 21\%/17\% with a 6\% area reduction over the industry-leading commercial synthesis tool.