Make Every Move Count: LLM-based High-Quality RTL Code Generation Using MCTS

TL;DR

This paper introduces a novel transformer decoding method using Monte Carlo tree search to improve the quality, correctness, and PPA efficiency of RTL code generated by large language models.

Contribution

The paper presents an automated decoding algorithm that incorporates MCTS to guide LLMs in producing more reliable and PPA-optimized RTL code, addressing previous limitations.

Findings

Consistently generates functionally correct RTL code.

Achieves 31.8% improvement in area-delay product for large designs.

Effectively addresses PPA-unawareness in LLM-based RTL generation.

Abstract

Existing large language models (LLMs) for register transfer level code generation face challenges like compilation failures and suboptimal power, performance, and area (PPA) efficiency. This is due to the lack of PPA awareness in conventional transformer decoding algorithms. In response, we present an automated transformer decoding algorithm that integrates Monte Carlo tree-search for lookahead, guiding the transformer to produce compilable, functionally correct, and PPA-optimized code. Empirical evaluation with a fine-tuned language model on RTL codesets shows that our proposed technique consistently generates functionally correct code compared to prompting-only methods and effectively addresses the PPA-unawareness drawback of naive large language models. For the largest design generated by the state-of-the-art LLM (16-bit adder), our technique can achieve a 31.8% improvement in the area-delay product.