2 * O(nlog n) work algorithm (Hillis-Steele)
6 void algorithm (numtype x[], unsigned long size, unsigned int ops[]) {
11 for(k=2; k <= size; k <<=1){
12 #pragma omp parallel for shared(x, size, ops, k) private(i)
13 for(i = (k-1); i < size; i+=k){
14 //printf ("x[%2li] = x[%2li] + x[%2li]; // {i:%li, k:%li}\n", i, i-k, i, i, k);
15 x[i] = x[i-(k/2)] + x[i];
17 //arrayState(x, size);
21 /* Copy out array to in */
22 // for (i = 0; i < size; i++){