Parallelize the reduced version of the n$-$body solver with OpenMP and a single critical directive to protect access to the forces array. Parallelize the rest of the solver by parallelizing the inner for loops. How does the performance of this code compare with the performance of the serial solver? Explain your answer.