This test checks for a defect found in the 4.0-4.0.2 releases of MPICH built
with the yaksa datatype engine.

The reduce operations for the fixed width unsigned integers (e.g. uint64_t,
a.k.a MPI_UINT64_T) are not performed correctly.

The bug is tracked here:

<https://github.com/pmodels/mpich/issues/6083>

Additionally, the test checks whether the reduce operations for other types that
are used in YAC are performed correctly.
