I can't find it help!!!

[fz6]

Hi,
I have a problem with this code:

#include <stdio.h>
#include <math.h>
#include <stdlib.h>

#define PI 3.14159265358979323846264338327

void func1(double *matrix1, int N_spaces);

int main(int argc,char **argv) {
double *matrix1=NULL;
int N_spaces = 16;
int i;

matrix1 = malloc((N_spaces+1)*2*sizeof(double));

func1(matrix1, N_spaces);
return 0;
}

void func1(double *matrix1, int N_spaces) {
double x,y,h;
int j,z;

h = 2/(double)N_spaces;

for(j=0,z=N_spaces+j; j<=N_spaces; j++,z++) {
x = -1+j*h;
y = sin(PI*x);
matrix1[j] = x;
matrix1[z] = y;


printf("%d %d %.7f %.7f\n", j, z, matrix1[j],matrix1[z] );
}

for(j=0,z=N_spaces+j; j<=N_spaces; j++,z++) {
printf("%d %d %.7f %.7f\n", j, z, matrix1[j],matrix1[z] );

}
}

The output is the following:
0 16 -1.0000000 -0.0000000
1 17 -0.8750000 -0.3826834
2 18 -0.7500000 -0.7071068
3 19 -0.6250000 -0.9238795
4 20 -0.5000000 -1.0000000
5 21 -0.3750000 -0.9238795
6 22 -0.2500000 -0.7071068
7 23 -0.1250000 -0.3826834
8 24 0.0000000 0.0000000
9 25 0.1250000 0.3826834
10 26 0.2500000 0.7071068
11 27 0.3750000 0.9238795
12 28 0.5000000 1.0000000
13 29 0.6250000 0.9238795
14 30 0.7500000 0.7071068
15 31 0.8750000 0.3826834
16 32 1.0000000 0.0000000

0 16 -1.0000000 1.0000000
1 17 -0.8750000 -0.3826834
2 18 -0.7500000 -0.7071068
3 19 -0.6250000 -0.9238795
4 20 -0.5000000 -1.0000000
5 21 -0.3750000 -0.9238795
6 22 -0.2500000 -0.7071068
7 23 -0.1250000 -0.3826834
8 24 0.0000000 0.0000000
9 25 0.1250000 0.3826834
10 26 0.2500000 0.7071068
11 27 0.3750000 0.9238795
12 28 0.5000000 1.0000000
13 29 0.6250000 0.9238795
14 30 0.7500000 0.7071068
15 31 0.8750000 0.3826834
16 32 1.0000000 0.0000000

As you notice the
0 16 -1.0000000 -0.0000000 and 0 16 -1.0000000 1.0000000 lines are different but they shouldnt!!!
Can you spot what I can't?
Thank you.

[Rubberman]

I haven't taken the time to analyze your code, but your #define for PI is not a valid double precision floating point number, which only allow approximately 15 significant digits of precision, but you are using 30. If your system supports long doubles then you can get 30 digits, but you aren't using that here. In any case, such a number can skew your results. Whether or not it will depends upon how the compiler deals with the extra digits. In any case, computer "innumeracy" of the sort you are experiencing is usually due to the programmer not understanding how digital processors handle numeric data like this, and so end up with invalid results upon occasion.

[fz6]

Hi,
thanx for the reply.
I tried reducing the digits of PI but the result was the same.

[nopycckn]

For one thing, at the end of main() - before return 0;:
free(matrix1);
is always the proper thing to do.
For another, why is:
int i;
declared in main()?
For another:
I get "undefined reference to sin".
sin() isn't in the headers you have in this code.
Last of all, I haven't spent too long looking at this, but I don't understand exactly what you're doing. Not to sound lazy, but if you wrote a description of exactly what you're trying to do it might help.

[fz6]

Hi,
thank you for the reply.
As for the free() before return I'd rather follow Theo De Raadt's advice and not to, as it is most likely to cause problems rather than solve them.

The undefined reference is due to you not passing the -lm flag to the compiler.

The program just uses a function that fills an array with values nothing more.

[fz6]

Check your compiler settings, ofcourse it compiles without errors.
Tip: cc prog.c -lm

[fz6]

cc is gcc
what errors do you get?

[fz6]

Thank you anyway
P.S. It really doesn't matter why I do it, it matters why it doesn't work

[fz6]

it can't get any simpler

[drl]

Hi.

Should your loop tests be "<=" or "<" ? ... cheers, drl