[R-sig-hpc] Rmpi and random number generation

[Hodgess, Erin]

I had tried the mpi.setup.rngstream but still obtained the same results.

________________________________________
From: Dirk Eddelbuettel [edd at debian.org]
Sent: Friday, July 19, 2013 4:58 PM
To: Hodgess, Erin
Cc: r-sig-hpc at r-project.org
Subject: Re: [R-sig-hpc] Rmpi and random number generation

On 19 July 2013 at 21:30, Hodgess, Erin wrote:
| Here is an Rmpi question, please:  I am trying to do a little demo program in which I generate a few random numbers.
[...]
| All of the random numbers are the same across the various variables.  Has anyone run into this, please?

1) See eg 'Section 6: Random-number generation' in the vignette of the
   package "parallel" (part of base R) for an introduction.

2) Even the most naive Google search for 'Rmpi parallel RNG' (and you surely
   tried something like that first, yes?) leads to this fine poster by the
   author of the Rmpi package:

      http://www.stats.uwo.ca/faculty/yu/Rmpi/download/rsr/poster.pdf

   which contains an example.

Dirk

--
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com