The MPI_Allgatherv operation each process collects data from
any other process and concats the received data in one buffer. In
additon to the MPI_Allgather operation, we can use per
processor receiving from another processes a specific displacement and
length, which determine where to write received data in the
root's buffer and how man bytes to receive from any processor. Of
course, it is interesting to see, what are the extra costs of this
features. We vary over the number of nodes. The messages have the
length of 256 Bytes (for each node).
Pattern: Collective varied over number of nodes.
default values: 8 nodes, message length 256 units, max. / act. time for suite disabled/0.01 min.
