Sphynx: A parallel multi-GPU graph partitioner for distributed-memory systems

Graph partitioning has been an important tool to partition the work among several processors to minimize the communication cost and balance the workload. While accelerator-based supercomputers are emerging to be the standard, the use of graph partitioning becomes even more important as applications...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Parallel computing Jg. 106; H. C; S. 102769
Hauptverfasser: Acer, Seher, Boman, Erik G., Glusa, Christian A., Rajamanickam, Sivasankaran
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Netherlands Elsevier B.V 01.09.2021
Elsevier
Schlagworte:
Distributed-memory systems
GPUs
Graph partitioning
Spectral partitioning
Graph partitioning
Distributed-memory systems
Spectral partitioning
GPUs
ISSN:0167-8191, 1872-7336
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Graph partitioning has been an important tool to partition the work among several processors to minimize the communication cost and balance the workload. While accelerator-based supercomputers are emerging to be the standard, the use of graph partitioning becomes even more important as applications are rapidly moving to these architectures. However, there is no distributed-memory-parallel, multi-GPU graph partitioner available for applications. We developed a spectral graph partitioner, Sphynx, using the portable, accelerator-friendly stack of the Trilinos framework. In Sphynx, we allow using different preconditioners and exploit their unique advantages. We use Sphynx to systematically evaluate the various algorithmic choices in spectral partitioning with a focus on the GPU performance. We perform those evaluations on two distinct classes of graphs: regular (such as meshes, matrices from finite element methods) and irregular (such as social networks and web graphs), and show that different settings and preconditioners are needed for these graph classes. The experimental results on the Summit supercomputer show that Sphynx is the fastest alternative on irregular graphs in an application-friendly setting and obtains a partitioning quality close to ParMETIS on regular graphs. When compared to nvGRAPH on a single GPU, Sphynx is faster and obtains better balance and better quality partitions. Sphynx provides a good and robust partitioning method across a wide range of graphs for applications looking for a GPU-based partitioner. •Sphynx is the first multi-GPU graph partitioner on distributed-memory systems.•Sphynx uses a spectral method followed by a fast geometric method.•Sphynx is flexible and provides multiple preconditioners in the eigensolver.•Sphynx has been tuned for different preconditioners and graph types.
Bibliographie:17-SC-20-SC; NA-0003525; AC05-00OR22725
USDOE National Nuclear Security Administration (NNSA)
ISSN:0167-8191
1872-7336
DOI:10.1016/j.parco.2021.102769