[Rd] Problem with R math library.

Dirk Eddelbuettel edd at debian.org
Thu Jan 28 18:23:47 CET 2010


On 28 January 2010 at 17:59, Guillaume Yziquel wrote:
| Guillaume Yziquel a écrit :
| > Dirk Eddelbuettel a écrit :
| >>
| >> Salut Guilluame,
| >>
| >> | > val norm_rand : unit -> float
| >> | > Random variates from the standard normal distribution. Bug: 
| >> currently systematically returns -8.77332116900134373.
| >> | | Any idea as to why the function systematically returns the same 
| >> value? | Is there a way the math library should be initialised?
| >>
| >> I think it is pretty clearly documented in R-exts:
| >>
| >>   However, before these are used, the user must call
| >>          GetRNGstate();
| >>     and after all the required variates have been generated, call
| >>          PutRNGstate();
| >>     These essentially read in (or create) `.Random.seed' and write it out
| >>   after use.
| > 
| > Fair enough. I admit I've been busy with low detail stuff, and omitted 
| > to come back to R-exts.
| > 
| > However, I have another question on which I do not find information (I 
| > found it once, but do not know how to find it again...): What's the big 
| > difference between using the R mathematical library in standalone mode 
| > and not in standalone mode? How does it translate in terms of C 
| > directives and linking modalities? I've noticed the MATHLIB_STANDALONE 
| > macro, but I do not know how I should use it...
| > 
| > All the best,
| 
| OK. So concerning the headers:
| 
| #define MATHLIB_STANDALONE
| #include <Rmath.h>
| 
| Concerning dependencies: compile against only libRmath.so. Not against 
| libR.so. (I wonder why this is so crucial, though...)

By design as libRmath is meanth to __standalone__ hence indepdent of R. This
is a a feature.  

This is probably not what _you_ want as you are embedding R, so you need libR.
 
| Concerning documentation: Section 9 The standalone Rmath library from 
| the R-admin.pdf documentation.
| 
| This solved my problem, since I'm able to generated random values with 
| norm_rand, unif_rand, etc...
| 
| Thank you for your help.

Always a pleasure.

Dirk

-- 
Three out of two people have difficulties with fractions.



More information about the R-devel mailing list