Addressing volume and latency overheads in 1d-parallel sparse matrix-vector multiplication

Date
2017-08-09
Advisor
Instructor
Source Title
Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing
Print ISSN
Electronic ISSN
Publisher
Springer
Volume
Issue
Pages
625 - 637
Language
English
Type
Conference Paper
Journal Title
Journal ISSN
Volume Title
Abstract

The scalability of sparse matrix-vector multiplication (SpMV) on distributed memory systems depends on multiple factors that involve different communication cost metrics. The irregular sparsity pattern of the coefficient matrix manifests itself as high bandwidth (total and/or maximum volume) and/or high latency (total and/or maximum message count) overhead. In this work, we propose a hypergraph partitioning model which combines two earlier models for one-dimensional partitioning, one addressing total and maximum volume, and the other one addressing total volume and total message count. Our model relies on the recursive bipartitioning paradigm and simultaneously addresses three cost metrics in a single partitioning phase in order to reduce volume and latency overheads. We demonstrate the validity of our model on a large dataset that contains more than 300 matrices. The results indicate that compared to the earlier models, our model significantly improves the scalability of SpMV. © 2017, Springer International Publishing AG.

Course
Other identifiers
Book Title
Keywords
Communication cost, Hypergraph partitioning, One-dimensional partitioning, Sparse matrix-vector multiplication
Citation
Published Version (Please cite this version)