[Rcpp-devel] Setting R seed from Cpp (continued)

Simon Zehnder szehnder at uni-bonn.de
Mon Jun 10 22:12:40 CEST 2013


Hi Dirk,

sorry for reposting. I thought this was specifically about getting the second approach of Matteo run in the same way as the first one.

Yeah, the RNG thing is something I myself haven't yet figured totally out. Especially in regard to Matteo's note, that the final task should run in parallel: I would strongly discourage his approach.

Usually a cycle of an RNG isn't exhausted in a simulation. So what one can do is to run a for loop in parallel and increase the seed in every iteration from 0 to (index* number of generated values). But for this one has 
to know exactly how much random values are produced. Otherwise it is rather undefined what happens. Here I would like to refer to another discussion on the list: http://www.mail-archive.com/rcpp-devel@lists.r-forge.r-project.org/msg04861.html, where Yan and you (next to Chris and Conrad) gave very 
good input. 


Best

Simon


On Jun 10, 2013, at 9:28 PM, Dirk Eddelbuettel <edd at debian.org> wrote:

> 
> Simon,
> 
> On 10 June 2013 at 20:22, Simon Zehnder wrote:
> | this is an interesting post. My assumption is, that as in the first approach
> 
> I fear tou may be a couple of posts late.  Please see
> 
>  http://thread.gmane.org/gmane.comp.lang.r.rcpp/5651
> 
> Thanks,  Dirk
> 
> -- 
> Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com



More information about the Rcpp-devel mailing list