Fusing Gathers with Integer Linear Programming

TL;DR

This paper introduces an Integer Linear Programming method to optimize fusion strategies in parallel programs, including complex gather operations, enhancing performance in combinator-based languages.

Contribution

It presents a novel ILP-based approach for optimal fusion in parallel programming, extending beyond traditional producer-consumer fusion to include gather operations.

Findings

Effective ILP-based fusion optimization for gather and other operations

Applicable to just-in-time compilation scenarios

Improves performance of combinator-based parallel programs

Abstract

We present an Integer Linear Programming based approach to finding the optimal fusion strategy for combinator-based parallel programs. While combinator-based languages or libraries provide a convenient interface for programming parallel hardware, fusing combinators to more complex operations is essential to achieve the desired performance. Our approach is not only suitable for languages with the usual map, fold, scan, indexing and scatter operations, but also gather operations, which access arrays in arbitrary order, and therefore goes beyond the traditional producer-consumer fusion. It can be parametrised with appropriate cost functions, and is fast enough to be suitable for just-in-time compilation.