[glug-t] g++ / fortran issue

Hi
 I have a cpp pgm which I am compiling like this
g++ try.cpp ode15s.f -lg2c

Contents of .cpp are somewhat this...
#define pi 3.1415926535
using namespace std;

int fcn_(int *neq, double *t, double *H, double *dH, double *rpar,
double 
*ipar);
extern "C" { void  ode15s_(int (*fcn_)(int *neq, double *t, double *H, 
double *dH, double *rpar, double *ipar), int *, double *, double *,
double 
(*)[101]); }

const int Nt=48, Nr=48;
const double Nr1=48,Nt1=48,dt=2*pi/Nt1,dr=1/Nr1;
double 
r[Nr+1][Nt+1],r1[(Nr+1)*(Nt+1)+1],r2[(Nr+1)*(Nt+1)],Th[Nr+1][Nt
+1],Th1[(Nr+1)*(Nt+1)+1],Th2[(Nr+1)*(Nt+1)],x[(Nt+1)*(Nr+1)],y[(Nr
+1)*(Nt+1)];

int main()
{
    double tend, atol;
    int i,j,k=0,k1,k2,neq;
        int des; //number of differential equations
        /*      int Nt=4,Nr=4; */
    double tout, t, H[Nr*Nt];
    ofstream outfile("nilesh.txt");
    neq = Nr*Nt;


2 matrices are used whose order is 48 x 48 , when i keep it 24 x 24 my 
compiled pgm runs well. When I change 24 to 48 then compiled pgm gives 
segmentation fault. on mdk/rh/fedora systems with RAM upto 2 GB and p4
HT 
processors.

What could be done to get rid of segmentation fault ? This is only when 
order of matrix change from 24 to 48

Thanking you,
upen
*******************************************************************************

---------------------------------------------------------------
To unsubscribe send a mail to glug_t-request@xxxxxxxxxxxxx with 
'unsubscribe' as subject.

Website: http://glugt.linuxisle.com

Other related posts: