| MPI_GRAPH_CREATE(comm_old, nnodes, index, edges, reorder, comm_graph) | |
| IN comm_old | input communicator (handle) |
| IN nnodes | number of nodes in graph (integer) |
| IN index | array of integers describing node degrees (see below) |
| IN edges | array of integers describing graph edges (see below) |
| IN reorder | ranks may be reordered ( true) or not ( false) (logical) |
| OUT comm_graph | new communicator with associated graph topology (handle) |
MPI_GRAPH_CREATE returns a handle to a new communicator to which the graph topology information is attached. If reorder = false then the rank of each MPI process in the group of the new communicator is identical to its rank in the group of the old communicator. If reorder = true then the procedure may reorder the ranks of the MPI processes. If the number of nodes in the graph ( nnodes) is smaller than the size of the group of comm_old, then MPI_COMM_NULL is returned by some MPI processes, in analogy to MPI_CART_CREATE and MPI_COMM_SPLIT. If the graph is empty, i.e., nnodes = 0, then MPI_COMM_NULL is returned in all MPI processes. The call is erroneous if it specifies a graph that is larger than the group size of the input communicator.
The three parameters nnodes, index and edges define the graph structure. nnodes is the number of nodes of the graph. The nodes are numbered from 0 to nnodes-1. The i-th entry of array index stores the total number of neighbors of the first i graph nodes. The lists of neighbors of nodes 0, 1, ..., nnodes-1 are stored in consecutive locations in array edges. The array edges is a flattened representation of the edge lists. The total number of entries in index is nnodes and the total number of entries in edges is equal to the number of graph edges.
The definitions of the arguments nnodes, index, and edges are illustrated with the following simple example.
Example
Specification of the adjacency matrix for MPI_GRAPH_CREATE.
Assume there are four MPI processes with ranks 0, 1, 2, 3 in the input communicator with the following adjacency matrix:
| MPI process | neighbors |
| 0 | 1, 3 |
| 1 | 0 |
| 2 | 3 |
| 3 | 0, 2 |
| nnodes = | 4 |
| index = | 2, 3, 4, 6 |
| edges = | 1, 3, 0, 3, 0, 2 |
Thus, in C, index[0] is the degree of node zero, and index[i] -
index[i-1] is the degree of node i, i=1, ..., nnodes-1;
the list of neighbors of node zero is stored in edges[j], for
and the
list of neighbors of node i,
,
is stored in edges[j], 
.
In Fortran, index(1) is the degree of node zero, and index(i+1) -
index(i) is the degree of node i, i=1, ..., nnodes-1;
the list of neighbors of node zero is stored in edges(j), for
and the list of neighbors of node
i, 
,
is stored in edges(j), 
.
A single MPI process is allowed to be defined multiple times in the list of
neighbors of an MPI process (i.e., there may be multiple edges between two
MPI processes). An MPI process is also allowed to be a neighbor to itself (i.e., a self
loop in the graph). The adjacency matrix is allowed to be nonsymmetric.
Advice to users.
Performance implications of using multiple edges or a nonsymmetric
adjacency matrix are not defined. The definition of a node-neighbor
edge does not imply a direction of the communication.
( End of advice to users.)
Advice
to implementors.
The following topology information is likely to be stored with a communicator: