MVAPICH2 vs. OpenMPI for a Clustering Algorithm

Thumbnail Image

Files

BlasbergGobbertHPCF20087.pdf (4.02 MB)

Links to Files

https://userpages.umbc.edu/~gobbert/papers/BlasbergGobbertHPCF20087.pdf

Permanent Link

http://hdl.handle.net/11603/11673

Collections

UMBC Mathematics and Statistics Department
UMBC Faculty Collection

Author/Creator

Author/Creator ORCID

Date

2008

Type of Work

technical report
Text

Department

Program

Citation of Original Publication

Rights

This item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.

Subjects

OpenMPI
Clustering Algorithm
MVAPICH2
memory-optimal
UMBC High Performance Computing Facility (HPCF)
affinity propogation

Abstract

Affinity propagation is a clustering algorithm that functions by identifying similar data points in an iterative process. Its structure allows for taking full advantage of parallel computing to enable the solution of larger problems and to solve them faster than possible in serial. We present a memory-optimal implementation with minimal number of communication commands and demonstrate its excellent scalability. Additionally, we present a comparison of two implementations of MPI that demonstrate that MVAPICH2 exhibits better scalability up to larger numbers of parallel processes than OpenMPI for this code.