A Scalable Distributed Parallel Breadth-First Search Algorithm on BlueGene/L

TLDR

Many emerging large‑scale data science applications require searching large graphs distributed across multiple memories and processors. This paper presents a distributed breadth‑first search (BFS) scheme that scales for random graphs with up to three billion vertices and 30 billion edges. The authors tested scalability on IBM BlueGene/L with 32,768 nodes, employing 2D edge partitioning, memory‑scalable optimizations, and efficient collective communication functions tailored to the 3D torus architecture. For Poisson random graphs, the expected message size scales with both 2D and 1D partitionings, and the algorithm’s performance and characteristics are measured and reported.

Abstract

Many emerging large-scale data science applications require searching large graphs distributed across multiple memories and processors. This paper presents a distributed breadth- first search (BFS) scheme that scales for random graphs with up to three billion vertices and 30 billion edges. Scalability was tested on IBM BlueGene/L with 32,768 nodes at the Lawrence Livermore National Laboratory. Scalability was obtained through a series of optimizations, in particular, those that ensure scalable use of memory. We use 2D (edge) partitioning of the graph instead of conventional 1D (vertex) partitioning to reduce communication overhead. For Poisson random graphs, we show that the expected size of the messages is scalable for both 2D and 1D partitionings. Finally, we have developed efficient collective communication functions for the 3D torus architecture of BlueGene/L that also take advantage of the structure in the problem. The performance and characteristics of the algorithm are measured and reported.