[Rd] Protocol for setting default number of cores

From: Udaya B. Kogalur <kogalurshear_at_gmail.com>
Date: Wed, 14 Nov 2012 16:22:15 -0500


We recently posted randomForestSRC on CRAN. It uses OpenMP in the native code extensively. We set the default number of cores to two (2), but we typically run it at the maximum (omp_get_max_threads()) during analysis. Currently, users need to set options(), set an environment variable, or edit their .Rprofile to use more than two cores. We followed the protocol for mc.cores in mclapply() in the package parallel. R CMD CHECK aside, is it acceptable to set the default to the maximum or is this not recommended?

ubk

-- 
kogalurshear_at_gmail.com (preferred)
udaya_at_kogalur-shear.com

Phone:     919-824-9825
Fax:         919-890-3868
Website:  www.kogalur-shear.com

Udaya B. Kogalur, Ph.D.
Adjunct Staff, Dept of Quantitative Health Sciences, Cleveland Clinic Foundation
Consultant, Kogalur Shear Corporation
6887 Idols Road, Suite B
Clemmons, NC 27012

______________________________________________
R-devel_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Received on Thu 15 Nov 2012 - 09:12:18 GMT

This quarter's messages: by month, or sorted: [ by date ] [ by thread ] [ by subject ] [ by author ]

All messages

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Fri 16 Nov 2012 - 08:00:51 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-devel. Please read the posting guide before posting to the list.

list of date sections of archive