[OT] statistical calculations (Martin WHEELER)

A.J. Rossini rossini at blindglobe.net
Mon Nov 24 15:04:55 EST 2003


Matthew Timothy Pratola <mp00aa at cosc.brocku.ca> writes:

> Anyhow, to keep this slightly on-topic, in a recent conversation with
> someone from the R project, i was told that there is maybe a rough,
> non-widely distributed implementation of MPI in R, which i think would be
> nice, but currently searching R and MPI on google does not yield much.
> Actually the person i spoke to gave me a name to search for, but i don't
> have that information in front of me right now...

Look for Rmpi.  I believe it's in the contrib non-current directory on
CRAN.  It works with LAM-MPI, though we've talked about extending it
to MPICH.

If interested in programming statistical calculations on a beowulf,
one might consider SNOW, which is an R library which provides a higher
(but simpler) level implementation (independent of PVM or MPI -- will
even use socket-based communication on a cluster if you don't have
it), and integrates transparently with SPRNG (the scalable parallel
RNG). 

See 

  http://www.analytics.washington.edu/~rossini/courses/cph-statcomp/

and Lecture/Lab 4 for description/issues in interactively computing
statistical quantities on a computational cluster (for statisticians
who don't want to figure out communication, and just want to get
results faster).

(I'm biased in my view -- we wrote the wrappers to PVM and SPRNG for
R, as well as contributed to SNOW, and just need to extend the current
set of MPI wrappers).

best,
-tony

-- 
rossini at u.washington.edu            http://www.analytics.washington.edu/ 
Biomedical and Health Informatics   University of Washington
Biostatistics, SCHARP/HVTN          Fred Hutchinson Cancer Research Center
UW (Tu/Th/F): 206-616-7630 FAX=206-543-3461 | Voicemail is unreliable
FHCRC  (M/W): 206-667-7025 FAX=206-667-4812 | use Email

CONFIDENTIALITY NOTICE: This e-mail message and any attachments may be
confidential and privileged. If you received this message in error,
please destroy it and notify the sender. Thank you.
_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list