quiz #5 question #5

[hughkf]

As indicated in the posted solution to question #5 in quiz #5, option #3 is wrong because it is inefficient, but otherwise gives the correct result. In that case, shouldn't the question have been "which of the following are correct and efficient parallelizations using OpenMP"?

[cswiercz]

The question asks

which of the following are correct parallelizations using OpenMP? (Select all that apply.)

I would argue that the code

void vec_add(double* out, double* v, double* w, int N)
{
  #pragma omp parallel
  {
    for (int i=0; i<N; ++i)
      out[i] = v[i] + w[i];
  }
}

is an incorrect parallelization because, although the code is run in parallel and uses the #pragma omp parallel directive in the correct way, the problem itself is not solved in parallel.