omitting a printf causes an incorrect answer - c

Question

For some reason I was getting a dramatically incorrect answer for a problem, so I put in this printf to try and debug.
for (s = 0; s <= 100; s++) {
for (t = 0; t <= 100; t++) {
printf("At (%f,%f), spl = %f\n", s, t, spl(loc_data, s, t)); */
if (spl(loc_data, s, t) > 80) {
p++;
}
}
}
If I omit the printf line, I get an incorrect answer. I think that it has something to do with memory allocation or uninitialised variables, but this is beyond my abilities. Any help would be appreciated.
Whole code:
#include <stdio.h>
#include <math.h>
typedef struct {
double x;
double y;
double W;
} data_t;
double spl(data_t *, double, double);
int main(int argc, char **argv) {
data_t loc_data[1000];
double spl0, p = 0, pp;
int i = 0, j = 0;
double s, t;
while (scanf("%lf %lf %lf", &loc_data[i].x, &loc_data[i].y, &loc_data[i].W) == 3) {
i++;
}
printf("\nStage 1\n=======\n");
printf("Number of sound sources: %d\n", i);
for (j = 0; j < i; j++) {
printf("%.1f meters east, %.1f meters north, power %1.5f Watts\n",
loc_data[j].x, loc_data[j].y, loc_data[j].W);
}
printf("\nStage 2\n=======\n");
spl0 = spl(loc_data, 0, 0);
printf("SPL at (0.0,0.0): %.1f dB\n", spl0);
printf("\nStage 3\n=======\n");
for (s = 0; s <= 100; s++) {
for (t = 0; t <= 100; t++) {
printf("At (%f,%f), spl = %f\n", s, t, spl(loc_data, s, t));
if (spl(loc_data, s, t) > 80) {
p++;
}
}
}
pp = p / 102.01;
printf("Points sampled: 10201\nAbove 80.0 dB: %.1f%%\n", pp);
return 0;
}
double spl(data_t *loc_data, double pointx, double pointy) {
int i = 0;
double r_sq, powi, spli, spl;
while (loc_data[i].W != 0) {
r_sq = pow(loc_data[i].x - pointx,2) + pow(loc_data[i].y - pointy,2);
powi = 10*log10(loc_data[i].W / pow(10,-12));
spli = powi + 10*log10((2 / (4 * M_PI * r_sq)) + (4 / (2.5 * M_PI * r_sq)));
spl = 10*log10(pow(10, spl/10) + pow(10, spli/10));
i++;
}
return spl;
}
Apologies for the poor formatting.

Answer 1

You''re using a variable (spl) before it's initialized inthe function spl():
double spl(data_t *loc_data, double pointx, double pointy) {
int i = 0;
double r_sq, powi, spli, spl;
while (loc_data[i].W != 0) {
r_sq = pow(loc_data[i].x - pointx,2) + pow(loc_data[i].y - pointy,2);
powi = 10*log10(loc_data[i].W / pow(10,-12));
spli = powi + 10*log10((2 / (4 * M_PI * r_sq)) + (4 / (2.5 * M_PI * r_sq)));
spl = 10*log10(pow(10, spl/10) + pow(10, spli/10));
// ^^^
i++;
}
return spl;
}
Calling printf() is probably influencing the value the variable happens to have.
Also, you read in a number of entries into the loc_data array, but don't pass that information to the spl() function. In spl() you treat the array entry with field W as the 'sentinel' - the end of the array. Is it a given that the last entry in the input data will have a zero value? If so, you should probably let us know, and probably check for that when done reading the input.

Answer 2

You are not initializing the loc_data entries in any way, which means they will contain garbage. After you scanf the entries (make sure to stop reading after you read 1000 lines), fill the remaining elements of the array with zeroes. I know you said you already tried this, but it's not visible in the code you posted.
Additionally, you should pass i (after giving it a better name) into the spl function so it knows how many good sources you have, instead of looking for a 0 as a terminator, which might not exist if you have 1000 good sources, or might exist too early if some source has a power of 0.