Doubling up would increase the error estimate because of the norm and thus would decrease the dt and cause the solution to be more accurate. If you use a lower solver tolerance the solution should be indistinguishable.
1 Like