xgrid and R: Parallel Distributed Processing Using Heterogeneous Groups of Apple Computers

Authors

Sarah C. Anoke, Smith College
Yuting Zhao, Smith College
Rafael Jaeger, Brown University
Nicholas J. Horton, Smith College

Date of this Version

6-2012

Document Type

Article

Citation

The R Journal (June 2012) 4(1)

Comments

Copyright 2012, The R Foundation. Open access material. License: CC BY 3.0 Unported

Abstract

The Apple Xgrid system provides access to groups (or grids) of computers that can be used to facilitate parallel processing. We describe the xgrid package which facilitates access to this system to undertake independent simulations or other long-running jobs that can be divided into replicate runs within R. Detailed examples are provided to demonstrate the interface, along with results from a simulation study of the performance gains using a variety of grids. Use of the grid for “embarassingly parallel” independent jobs has the potential for major speedups in time to completion. Appendices provide guidance on setting up the workflow, utilizing add on packages, and constructing grids using existing machines.