Optimizing LLVM Pass Sequences with Shackleton: A Linear Genetic Programming Framework

TL;DR

This paper presents Shackleton, a linear genetic programming framework, for automatically optimizing sequences of LLVM compiler passes to improve software performance, demonstrating its effectiveness on different application complexities.

Contribution

Introduces Shackleton, a novel framework applying linear genetic programming to optimize LLVM pass sequences, advancing automated compiler optimization techniques.

Findings

Shackleton effectively optimizes LLVM pass sequences for different software complexities.

Hyperparameter analysis reveals key settings for optimization performance.

Automatic optimization can surpass handcrafted pass sequences.

Abstract

In this paper we introduce Shackleton as a generalized framework enabling the application of linear genetic programming -- a technique under the umbrella of evolutionary algorithms -- to a variety of use cases. We also explore here a novel application for this class of methods: optimizing sequences of LLVM optimization passes. The algorithm underpinning Shackleton is discussed, with an emphasis on the effects of different features unique to the framework when applied to LLVM pass sequences. Combined with analysis of different hyperparameter settings, we report the results on automatically optimizing pass sequences using Shackleton for two software applications at differing complexity levels. Finally, we reflect on the advantages and limitations of our current implementation and lay out a path for further improvements. These improvements aim to surpass hand-crafted solutions with an automatic discovery method for an optimal pass sequence.