PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

Issue with booleen when linking Fortran to C++ code

 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling
View previous topic :: View next topic  
Author Message
xlapillonne



Joined: 16 Feb 2011
Posts: 69
PostPosted: Wed Mar 13, 2013 11:50 am    Post subject: Issue with booleen when linking Fortran to C++ code Reply with quote
Hi,

I am trying to link some Fortran code to some C++ code which was compiled with gcc. The problem is that when passing the booleen (l1) to the C++ part, true values are associated to 255. This is an issue when evaluating ( ! l1), as it is then 254 which is still true. I have made a small example to illustrate the problem:

Main.f90
Code:

PROGRAM test
  USE, INTRINSIC :: iso_c_binding
  REAL*8 :: x(2)
  LOGICAL :: l1,l2
  LOGICAL(KIND=C_BOOL) :: cl1,cl2

  ! external functions
  INTERFACE
     SUBROUTINE cfunc(a,l1,l2) &
          BIND(c, name='cfunc_')
       USE, INTRINSIC :: iso_c_binding
       REAL*8 :: a(:)   
       LOGICAL(KIND=C_BOOL), value :: l1,l2

     END SUBROUTINE cfunc
  END INTERFACE
 
  x(1)=3.0
  x(2)=4.0

  l1=.TRUE.
  l2=.FALSE.

  cl1=l1
  cl2=l2
  call cfunc(x,cl1,cl2)

END PROGRAM test


cfunc.cpp :
Code:

#include <iostream>     // std::cout
#include "cfunc.h"

void cfunc_(double* a, bool l1, bool l2)
{
  bool ltest;

  ltest=255;

  std::cout<< "out: " <<  a[0] << " " << a[1]  << " " << l1 << " " << ! l1 << " " << ltest << " " << ! ltest << " "<< std::endl;

  if (! l1) std::cout<< "l1: Error (not True) is (True) !" << std::endl;
  if (! ltest) std::cout<< "ltest: Error (not True) is (True) !" << std::endl;

 
}


cfunc.h :
Code:

extern "C" {

  void cfunc_(double* a, bool l1, bool l2);

}


If I compile and run this I get:
Code:

+ g++ -c cfunc.cpp
+ pgf90 -lstdc++ cfunc.o -o test Main.f90
Main.f90:
test/C_Fortran> ./test
out: 3 4 255 254 1 0
l1: Error (not True) is (True) !


I though that using iso_c_binding would help, but apparenlty it's not. Do you have any idea/suggestions on how I could preceed ? (Note that if I compile with gfortran I don't have this issue)

Thanks,

Xavier

PS : I have been using PGI 13.2
Back to top
View user's profile
Deborah Caruso



Joined: 20 Oct 2004
Posts: 43
PostPosted: Fri Mar 15, 2013 12:06 pm    Post subject: Reply with quote
The issue here is that in pgf90 we set .TRUE. to 0xff ,
while gfortran sets .TRUE. to 1.

Use the compiler flag pgf90 -Munixlogical to use the unix logical value of .TRUE.
Back to top
View user's profile
xlapillonne



Joined: 16 Feb 2011
Posts: 69
PostPosted: Mon Mar 18, 2013 8:23 am    Post subject: Reply with quote
thanks !
Back to top
View user's profile
ofuhrer



Joined: 18 Feb 2008
Posts: 16
PostPosted: Mon Mar 18, 2013 8:36 am    Post subject: Other required flags? Reply with quote
Hi,

Thanks Deborah for the answer. Are there any other flags which need to be set in order to ensure PGI/GNU interoperability?

Cheers,
Oli
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 4996
Location: The Portland Group Inc.
PostPosted: Mon Mar 18, 2013 11:04 am    Post subject: Reply with quote
Hi Oli,

Other than the normal C++/Fortran interoperability issues, I don't believe that there are further GNU specific issues. Possible, but we're not aware of them.

Note that we did just add a PGI C++ compiler, pgc++, which is interoperable with g++.

- Mat
Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling All times are GMT - 7 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2002 phpBB Group