3 * Non-recursive, data parallel implementation
6 void algorithm (numtype x[], unsigned long size, unsigned int ops[]) {
11 for(k=1; k < size; k = kk){
13 for(i = kk-1; i < size; i+= kk){
19 for(k=kk>>1; k > 1; k = kk){
21 for(i = k-1; i < size-kk; i+= k){
22 x[i+kk] = x[i] + x[i+kk];