Leveraging Large Language Models for Generalizing Peephole Optimizations

TL;DR

This paper introduces LPG, a framework that leverages large language models to automate and improve the generalization of peephole optimizations in compilers, achieving higher success rates than existing methods.

Contribution

LPG integrates LLMs with formal validation to effectively generalize peephole optimizations, reducing manual effort and surpassing prior automation approaches like Hydra.

Findings

LPG successfully generalizes 90 out of 102 optimizations from LLVM.

LPG outperforms Hydra, generalizing 74 out of 81 optimizations compared to 35.

The framework combines semantic abstraction with formal validation for reliable optimization generalization.

Abstract

Peephole optimizations are a core component of modern optimizing compilers. It rewrites specific instruction into semantically equivalent but more efficient forms. In practice, creating a new peephole optimization often starts from a concrete optimization instance and requires lifting it into a more general rewrite rule that matches a wider range of instruction patterns. This generalization step is critical to optimization effectiveness, but it is also difficult: producing rules that are both correct and sufficiently general typically demands substantial manual effort and domain expertise. Existing approaches such as Hydra attempt to automate this task with program synthesis, but their generalization capability is often limited by search-space explosion, under-generalization, and restricted support for diverse instruction domains. We present LPG, large language model aided peephole optimization generalization, a framework that uses large language models (LLMs) to generalize peephole optimizations. The design of LPG is motivated by the observation that LLMs are effective at semantic abstraction and exploratory reasoning, while formal analyses are necessary to ensure that generated rules are sound and profitable. Based on this observation, LPG adopts a closed-loop workflow that integrates LLM-driven symbolic constant generalization, structural generalization, constraint relaxation, and bitwidth/precision generalization with feedback from syntactic validation, semantic verification, and profitability checking. We evaluate LPG on real-world peephole optimization issues drawn from the LLVM ecosystem. Overall, LPG successfully generalizes 90 out of 102 optimizations. On the integer-focused subset that is directly comparable to Hydra, LPG generalizes 74 out of 81 optimizations, whereas Hydra generalizes 35.