Openmp segmentation fault when setting the number of threads through console - c

Question

I have the code of matrices multiplication with using openmp:
#include <stdio.h>
#include <omp.h>
#include <math.h>
#define N 1000
int main()
{
long int i, j, k;
//long int N = atoi(argv[1]);
double t1, t2;
double a[N][N],b[N][N],c[N][N];
for (i=0; i<N; i++)
for (j=0; j<N; j++)
a[i][j]=b[i][j]=log(i*j/(i*j+1.)+1) +exp(-(i+j)*(i+j+1.));
t1=omp_get_wtime();
#pragma omp parallel for shared(a, b, c) private(i, j, k)
for(i=0; i<N; i++){
for(j=0; j<N; j++){
c[i][j] = 0.0;
for(k=0; k<N; k++) c[i][j]+=a[i][k]*b[k][j];
}
}
t2=omp_get_wtime();
printf("Time=%lf\n", t2-t1);
}
Now I want to set the number of threads which I want through command line. I do that by using
atoi(argv[])
Namely
#include <stdio.h>
#include <omp.h>
#include <math.h>
#define N 1000
int main(int argc, char** argv[])
{
long int i, j, k;
//long int N = atoi(argv[1]);
double t1, t2;
double a[N][N],b[N][N],c[N][N];
for (i=0; i<N; i++)
for (j=0; j<N; j++)
a[i][j]=b[i][j]=log(i*j/(i*j+1.)+1) +exp(-(i+j)*(i+j+1.));
int t = atoi(argv[1]);
t1=omp_get_wtime();
#pragma omp parallel for shared(a, b, c) private(i, j, k) num_threads(t)
for(i=0; i<N; i++){
for(j=0; j<N; j++){
c[i][j] = 0.0;
for(k=0; k<N; k++) c[i][j]+=a[i][k]*b[k][j];
}
}
t2=omp_get_wtime();
printf("Time=%lf\n", t2-t1);
}
Everything is fine, except one crucial thing: when I try to compute the product of matrices with dimension more than (more or less) 500, I get the mistake: "segmentation fault". Could someone clarify the reason for this mistake?

Answer 1

I don't know anything about openmp, but you are most assuredly blowing up your stack. Default stack space will vary from system to system, but with N == 1000, you are trying to put three 2D arrays totaling 3 million doubles on the stack. Assuming a double is 8 bytes, that's 24 million bytes, or just shy of 22.9MB. There can't be many systems allowing that kind of stack space. Instead, I'd recommend trying to grab that amount of memory from the heap. Something like this:
//double a[N][N],b[N][N],c[N][N];
double **a, **b, **c;
a = malloc(sizeof(double*) * N);
b = malloc(sizeof(double*) * N);
c = malloc(sizeof(double*) * N);
for (i=0; i<N; i++)
{
a[i] = malloc(sizeof(double) * N);
b[i] = malloc(sizeof(double) * N);
c[i] = malloc(sizeof(double) * N);
}
// do your calculations
for (i=0; i<N; i++)
{
free(a[i]);
free(b[i]);
free(c[i]);
}
free(a);
free(b);
free(c);
I've verified on my machine at least, that with N == 1000 I crash right out of the gate with EXC_BAD_ACCESS when trying to place those arrays on the stack. When I dynamically allocate the memory instead as shown above, I get no seg faults.