POET: Power-Oriented Evolutionary Tuning for LLM-Based RTL PPA Optimization

TL;DR

POET is a framework that uses evolutionary algorithms and LLMs to optimize RTL designs for power efficiency while ensuring correctness, eliminating hallucinations, and effectively navigating the PPA trade-off space.

Contribution

POET introduces a novel LLM-driven evolutionary tuning framework with a differential-testing approach for correctness and power-first search for PPA optimization.

Findings

Achieves 100% functional correctness across all tested designs.

Attains the best power reduction on all 40 RTL designs.

Provides competitive improvements in area and delay.

Abstract

Applying large language models (LLMs) to RTL code optimization for improved power, performance, and area (PPA) faces two key challenges: ensuring functional correctness of optimized designs despite LLM hallucination, and systematically prioritizing power reduction within the multi-objective PPA trade-off space. We propose POET (Power-Oriented Evolutionary Tuning), a framework that addresses both challenges. For functional correctness, POET introduces a differential-testing-based testbench generation pipeline that treats the original design as a functional oracle, using deterministic simulation to produce golden references and eliminating LLM hallucination from the verification process. For PPA optimization, POET employs an LLM-driven evolutionary mechanism with non-dominated sorting, power-first intra-level ranking, and proportional survivor selection to steer the search toward the low-power region of the Pareto front without manual weight tuning. Evaluated on the RTL-OPT benchmark across 40 diverse RTL designs, POET achieves 100% functional correctness, the best power on all 40 designs, and competitive area and delay improvements.