Rewrite System Showdown: Stochastic Search vs. EqSat

2026-05-18

AI Summary

This study addresses the lack of systematic comparison between mainstream approaches to equational program optimization—such as e-graph-based equality saturation—and alternative strategies like random search. For the first time, the authors conduct a rigorous empirical evaluation of both methods under a unified experimental setup across five standard benchmarks. Their findings elucidate the practical advantages of e-graphs in terms of optimization efficacy as well as their limitations and applicable boundaries. This work fills a critical gap in methodological assessment within the field and establishes a reliable baseline for future research, providing strong empirical grounding for comparative studies in equational reasoning and program optimization.

Abstract

Equality saturation has become a dominant paradigm for equational program optimization. However, it has never been rigorously compared to another approach to the same problem, even though several exist, the most notable being stochastic search. In this paper, we compare equality saturation to stochastic search over five benchmarks to answer the question: are e-graphs actually good?

Problem

• equality saturation
• stochastic search
• program optimization
• e-graphs
• rewrite systems

Innovation

• equality saturation
• stochastic search
• e-graphs
• program optimization
• empirical comparison