Integration using pointer

[shallowbay]

So I was playing around with trying to do an integral using pointers. I've written what I'm trying to do already using arrays

#include <stdio.h>
#include <math.h>
#define RANGE 500 // Final temp = T0 + RANGE


int main()
{
    int i, j, h = 1, T0 = 300;
    float k300 = 148, alpha = -1.25;   
    float Temp[RANGE + 1], k[RANGE + 1], theta[RANGE];   
    float sum[RANGE], trap[RANGE - 1];        
    
    for (i = 0; i <= RANGE ; i++) 
    {
        Temp[i] = i + T0;                                     
        k[i] = k300 * pow((Temp[i]/T0), alpha);  
    }        
     
    for (j = 0; j <= RANGE; j++) 
    {
        trap[j] = 0.5*h*(k[j+1] + k[j]);                    // Trapezoidal Integration
        sum[j] = sum[j-1] + trap[j];                            
        if (j > 0) 
        {
          theta[j] = T0 + (1/k300)*sum[j-1];  
        }       
    }
    
    return 0;
}

which works just fine. Instead I want to try to go through it using pointers. Is there a good way to go through the first loop above, something like?

    #include <stdio.h>
    #include <math.h>
    #define N 501 // Final temp = T0 + N
    
    
int main()
{

    int T0 = 300;
    float k300 = 148, alpha = -1.25;         
    float temp[N], kappa[N], theta[N -1];   
    float sum, trap[N - 1];  
    float *tl_i = temp, *tl_start = temp, *tl_end = (temp + N);   
    float *kappa_i = kappa, *kappa_end = (kappa + N), *kappa_start = kappa;
    float *trap_i = trap, *trap_start = trap, *trap_end = (trap + N - 1);
    float *kirch = theta;
    

    for (tl_i; tl_i<tl_end ; ++tl_i) 
    {
        *tl_i = (tl_i - tl_start) + T0;            
        *kappa_i = k300 * pow((*tl_i/T0), alpha);     
    }

[Salem]

> float Temp[RANGE + 1], k[RANGE + 1], theta[RANGE];
> float sum[RANGE], trap[RANGE - 1];
All these +/-1 on your array sizes are an invitation to buffer overflows.

> for (j = 0; j <= RANGE; j++)
> trap[j] = 0.5*h*(k[j+1] + k[j]); // Trapezoidal Integration
The last TWO indices of trap are off in the weeds somewhere.

We'll discuss pointers when you can demonstrate bug-free (at least as far as subscripting goes) array code.

[shallowbay]

Thanks for the help! I believe that this is a little better:

#include <stdio.h>
#include <math.h>
#define RANGE 501 // Final temp = T0 + RANGE


int main()
{
    int i, j, h = 1, T0 = 300;
    float k300 = 148, alpha = -1.25;  
    float Temp[RANGE], k[RANGE], theta[RANGE];    
    float sum[RANGE], trap[RANGE];              
    
    for (i = 0; i < RANGE ; i++) {
        Temp[i] = i + T0;                           
        k[i] = k300 * pow((Temp[i]/T0), alpha);        
    }        
     
    for (j = 0; j < RANGE; j++) {
        trap[j] = 0.5*h*(k[j+1] + k[j]);                    // Trapezoidal Integration
        sum[j] = sum[j-1] + trap[j];                            
        if (j > 0) {
          theta[j] = T0 + (1/k300)*sum[j-1]; 
        }       
    }
    
    return 0;
}

[Salem]

> trap[j] = 0.5*h*(k[j+1] + k[j]); // Trapezoidal Integration
Still an issue with the j+1 on the last iteration of the loop.

> sum[j] = sum[j-1] + trap[j];
And here on the first iteration.

[shallowbay]

> trap[j] = 0.5*h*(k[j+1] + k[j]); // Trapezoidal Integration
Still an issue with the j+1 on the last iteration of the loop.

> sum[j] = sum[j-1] + trap[j];
And here on the first iteration.

I used printf statements to try to make sure everything was right. Results seemed good, so let's hope for the best.....

#include <stdio.h>
#include <math.h>
#define RANGE 501 // Final temp = T0 + RANGE


int main()
{
    int i, j, h = 1, T0 = 300;
    float k300 = 148, alpha = -1.25;  
    float Temp[RANGE], k[RANGE], theta[RANGE];    
    float sum[RANGE], trap[RANGE];              
    
    for (i = 0; i < RANGE ; i++) {
        Temp[i] = i + T0;                           
        k[i] = k300 * pow((Temp[i]/T0), alpha);        
    }        
     
    for (j = 1; j < RANGE; j++) {
        sum[0] = 0;
        trap[j] = 0.5*h*(k[j] + k[j-1]);                    // Trapezoidal Integration
        sum[j] = sum[j-1] + trap[j];                           
        theta[j] = T0 + (1/k300)*sum[j];
        printf("%f    %f\n", theta[j], Temp[j]);
    }
    
    return 0;
}

[Salem]

To answer your original question, yes you can always express arrays using pointer notation, but it is seldom worth the effort.

So
int a[N];
for ( i = 0 ; i < N ; i++ ) a[i] = 0;

becomes
end = a + N;
for ( p = a ; p < end ; p++ ) *p = 0;

[Salem]

To answer your original question, yes you can always express arrays using pointer notation, but it is seldom worth the effort.

So
int a[N];
for ( i = 0 ; i < N ; i++ ) a[i] = 0;

becomes
end = a + N;
for ( p = a ; p < end ; p++ ) *p = 0;