Question

I have the following nested for loops. I want to parallelize the first loop but the second loop must NOT be parallelized. So I want that each thread executes the second loop in itself, which means that second loop must be special for each thread (for each “i” in the code).

How can I do that?

#pragma omp parallel for
    for (i=k+1;i<row;i++){
        for (n=0;n<k;n++){
           // #pragma omp atomic
            dummy += L[i][n]*L[k][n];
            L[i][k] = (A[i][k] - dummy)/L[k][k];
        }
        dummy = 0;
    }

Answer 1

The omp parallel for pragma applies only to the loop that follows it immediately. The i loop will be distributed between threads. Within each iteration of i, the n loop will be executed on the same thread. Your code already does what you are trying to do.