This "distribution" actually redundantly stores every entry of the associated matrix on every process. Again using a 2 \times 3 process grid, the entries of a 7 \times 7 matrix would be owned by the following sets of processes:
\left(\begin{array}{ccccccc} \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \\ \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} & \{0,1,...,5\} \end{array}\right)
.. cpp:class:: DistMatrix<scalarType,STAR,STAR>
.. cpp:class:: DistMatrix<scalarType,STAR,STAR,ELEMENT>