PROGRAML: A Graph-based Program Representation for Data Flow Analysis and Compiler Optimizations

Machine learning (ML) is increasingly seen as a viable approach for building compiler optimization heuristics, but many ML methods cannot replicate even the simplest of the data flow analyses that are critical to making good optimization decisions. We posit that if ML cannot do that, then it is insufficiently able to reason about programs. We formulate data flow analyses as supervised learning tasks and introduce a large open dataset of programs and their corresponding labels from several analyses. We use this dataset to benchmark ML methods and show that they struggle on these fundamental program reasoning tasks. We propose PROGRAML - Program Graphs for Machine Learning - a language-independent, portable representation of program semantics. PROGRAML overcomes the limitations of prior works and yields improved performance on downstream optimization tasks.

Automated summary

The paper discusses the application of machine learning in building compiler optimization heuristics, highlighting the challenges many machine learning methods face in program data flow analysis. By formulating data flow analysis as supervised learning tasks and introducing a dataset of programs and corresponding labels, the authors propose a language-independent representation of program semantics, PROGRAML, which overcomes limitations of prior works and improves performance on optimization tasks.