MPI notebook uses time_order=0

[FabioLuporini]

As explained in #981 , time_order=0 may sequentialize otherwise parallel loops. The MPI overview notebook creates TimeFunctions with time_order=0, which is therefore bad.

The only reason the notebook produces sane numbers is the small grid shape and the fact that DEVITO_MPI=basic is used (with DEVITO_MPI=full it'd fail for example), which completes the halo exchanges before computation