[Rd] Problem with R math library. [and a minor bug report]

[Dirk Eddelbuettel]

On 28 January 2010 at 14:00, Guillaume Yziquel wrote:
| 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...

Well, Debian comes to the rescue. Do 'sudo apt-get install r-mathlib' to the
standalone library. Then copy the example file out, add the missing 'extern
N01type N01_kind;' (and that it is missing may well be a bug -- I used 2.10.1
here -- anyone from R Core listening in who can add the one line ?) 

   edd at ron:~> cp  /usr/share/doc/r-mathlib/examples/test.c /tmp/
   edd at ron:~> grep extern /tmp/test.c
   extern N01type N01_kind; /* from nmath/snorm.c */
   edd at ron:~> gcc  -o /tmp/mathlibtest  /tmp/test.c -lRmath -lm
   edd at ron:~> /tmp/mathlibtest 
   *** loaded '/tmp/mathlibtest'
   one normal 1.119638
   normal via BM -1.734578
   edd at ron:~> 

I hope this answers your questions -- the standalone math library links with
only its libRmath and nothing else from R. It also only consume Rmath.h
(which I put into /usr/include so you need no -I arguments to gcc).

Hope this helps,  Dirk

PS For reference, here is test.c with line edited in:

/*
 *  Mathlib : A C Library of Special Functions
 *  Copyright (C) 2000-7  The R Development Core Team
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, a copy is available at
 *  http://www.r-project.org/Licenses/
 *
 */

#define MATHLIB_STANDALONE 1
#include <Rmath.h>

#include <stdio.h>
typedef enum {
    BUGGY_KINDERMAN_RAMAGE,
    AHRENS_DIETER,
    BOX_MULLER,
    USER_NORM,
    INVERSION,
    KINDERMAN_RAMAGE
} N01type;

extern N01type N01_kind; /* from nmath/snorm.c */
 
int
main(int argc, char** argv)
{
/* something to force the library to be included */
    qnorm(0.7, 0.0, 1.0, 0, 0);
    printf("*** loaded '%s'\n", argv[0]);
    set_seed(123, 456);
    N01_kind = AHRENS_DIETER;
    printf("one normal %f\n", norm_rand());
    set_seed(123, 456);
    N01_kind = BOX_MULLER;
    printf("normal via BM %f\n", norm_rand());
    
    return 0;
}


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