Bine Trees: Enhancing Collective Operations by Optimizing Communication Locality

TL;DR

Bine trees are a new family of collective algorithms designed to optimize communication locality, significantly reducing global-link traffic and improving performance on large HPC systems with various network topologies.

Contribution

Introduction of Bine trees, a novel collective algorithm family that enhances communication locality and reduces global-link traffic in HPC systems.

Findings

Up to 33% reduction in global-link traffic.

Achieved up to 5x speedups on large-scale supercomputers.

Consistent performance improvements across different topologies.

Abstract

Communication locality plays a key role in the performance of collective operations on large HPC systems, especially on oversubscribed networks where groups of nodes are fully connected internally but sparsely linked through global connections. We present Bine (binomial negabinary) trees, a family of collective algorithms that improve communication locality. Bine trees maintain the generality of binomial trees and butterflies while cutting global-link traffic by up to 33%. We implement eight Bine-based collectives and evaluate them on four large-scale supercomputers with Dragonfly, Dragonfly+, oversubscribed fat-tree, and torus topologies, achieving up to 5x speedups and consistent reductions in global-link traffic across different vector sizes and node counts.