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

int main(void) {

	int N = 1024, i, j, k;
	double **a, **b, **c;
	a = (double **)malloc(N*sizeof(double *));
	b = (double **)malloc(N*sizeof(double *));
	c = (double **)malloc(N*sizeof(double *));

	for (i = 0; i < N; i++){
		a[i] = (double *)malloc(N*sizeof(double));
		b[i] = (double *)malloc(N*sizeof(double));
		c[i] = (double *)malloc(N*sizeof(double));
	}

	for (i = 0; i < N; i++){
		for (j = 0; j < N; j++){
			a[i][j] = 5;
			b[i][j] = 5;
		}
	}

	for(i = 0; i < N; i++){
	  double *orig_pa = &a[i][0];
	  for(j = 0; j < N; j++){
		double *pa = orig_pa;
		double *pb = &b[0][j];
		register double suma = 0;
		for(k = 0; k < N; k++){
		  suma += *pa * *pb;
		  pa++;
		  pb += N;
		}
		c[i][j] = suma;
	  }
	}
	return 0;
}