If a topology has been defined with one of the above functions, then the topology information can be looked up using inquiry functions. They all are local calls.
| MPI_TOPO_TEST(comm, status) | |
| IN comm | communicator (handle) |
| OUT status | topology type of communicator comm (state) |
int MPI_Topo_test(MPI_Comm comm, int *status)
MPI_TOPO_TEST(COMM, STATUS, IERROR)
The output value status is one of the following:
int MPI_Graphdims_get(MPI_Comm comm, int *nnodes, int *nedges)
MPI_GRAPHDIMS_GET(COMM, NNODES, NEDGES, IERROR)
The information provided by MPI_GRAPHDIMS_GET can be used
to dimension the
vectors index and edges correctly for the following call
to MPI_GRAPH_GET.
int MPI_Graph_get(MPI_Comm comm, int maxindex, int maxedges, int *index, int *edges)
MPI_GRAPH_GET(COMM, MAXINDEX, MAXEDGES, INDEX, EDGES, IERROR)
int MPI_Cartdim_get(MPI_Comm comm, int *ndims)
MPI_CARTDIM_GET(COMM, NDIMS, IERROR)
int MPI_Cart_get(MPI_Comm comm, int maxdims, int *dims, int *periods, int *coords)
MPI_CART_GET(COMM, MAXDIMS, DIMS, PERIODS, COORDS, IERROR)
int MPI_Cart_rank(MPI_Comm comm, int *coords, int *rank)
MPI_CART_RANK(COMM, COORDS, RANK, IERROR)
For dimension i with periods(i) = true, if the coordinate,
coords(i), is out of range, that is, coords(i) < 0 or
coords(i)
If comm is associated with a zero-dimensional Cartesian topology,
coords is not significant and 0 is returned in rank.
int MPI_Cart_coords(MPI_Comm comm, int rank, int maxdims, int *coords)
MPI_CART_COORDS(COMM, RANK, MAXDIMS, COORDS, IERROR)
int MPI_Graph_neighbors_count(MPI_Comm comm, int rank, int *nneighbors)
MPI_GRAPH_NEIGHBORS_COUNT(COMM, RANK, NNEIGHBORS, IERROR)
int MPI_Graph_neighbors(MPI_Comm comm, int rank, int maxneighbors, int *neighbors)
MPI_GRAPH_NEIGHBORS(COMM, RANK, MAXNEIGHBORS, NEIGHBORS, IERROR)
Thus, the input arguments to MPI_GRAPH_CREATE are:
Therefore, calling MPI_GRAPH_NEIGHBORS_COUNT and
MPI_GRAPH_NEIGHBORS for each of the 4 processes will return:
INTEGER COMM, STATUS, IERROR
{ int MPI::Comm::Get_topology() const (binding deprecated, see Section Deprecated since MPI-2.2
) }
The function MPI_TOPO_TEST returns the type of topology that
is assigned to a communicator.
MPI_GRAPHDIMS_GET(comm, nnodes, nedges) IN comm communicator for group with graph structure (handle) OUT nnodes number of nodes in graph (integer) (same as number of processes in the group) OUT nedges number of edges in graph (integer)
INTEGER COMM, NNODES, NEDGES, IERROR
{ void MPI::Graphcomm::Get_dims(int nnodes[], int nedges[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
Functions MPI_GRAPHDIMS_GET and
MPI_GRAPH_GET retrieve the graph-topology information
that was associated with a communicator by
MPI_GRAPH_CREATE.
MPI_GRAPH_GET(comm, maxindex, maxedges, index, edges) IN comm communicator with graph structure (handle) IN maxindex length of vector index in the calling program
(integer) IN maxedges length of vector edges in the calling program
(integer) OUT index array of integers containing the graph structure (for details see the definition of MPI_GRAPH_CREATE) OUT edges array of integers containing the graph structure
INTEGER COMM, MAXINDEX, MAXEDGES, INDEX(*), EDGES(*), IERROR
{ void MPI::Graphcomm::Get_topo(int maxindex, int maxedges, int index[], int edges[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
MPI_CARTDIM_GET(comm, ndims) IN comm communicator with Cartesian structure (handle) OUT ndims number of dimensions of the Cartesian structure (integer)
INTEGER COMM, NDIMS, IERROR
{ int MPI::Cartcomm::Get_dim() const (binding deprecated, see Section Deprecated since MPI-2.2
) }
The functions MPI_CARTDIM_GET and
MPI_CART_GET return the Cartesian topology information that was
associated with a communicator by
MPI_CART_CREATE.
If comm is associated with a zero-dimensional Cartesian topology,
MPI_CARTDIM_GET returns ndims=0 and MPI_CART_GET will keep
all output arguments unchanged.
MPI_CART_GET(comm, maxdims, dims, periods, coords) IN comm communicator with Cartesian structure (handle) IN maxdims length of vectors dims, periods, and coords in the calling program (integer) OUT dims number of processes for each Cartesian dimension (array of integer) OUT periods periodicity ( true/ false) for each Cartesian dimension (array of logical) OUT coords coordinates of calling process in Cartesian structure (array of integer)
INTEGER COMM, MAXDIMS, DIMS(*), COORDS(*), IERROR
LOGICAL PERIODS(*)
{ void MPI::Cartcomm::Get_topo(int maxdims, int dims[], bool periods[], int coords[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
MPI_CART_RANK(comm, coords, rank) IN comm communicator with Cartesian structure (handle) IN coords integer array (of size ndims) specifying the Cartesian coordinates of a process OUT rank rank of specified process (integer)
INTEGER COMM, COORDS(*), RANK, IERROR
{ int MPI::Cartcomm::Get_cart_rank(const int coords[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
For a process group with Cartesian structure, the function
MPI_CART_RANK translates the logical process coordinates to process
ranks as they are used by the point-to-point routines.
dims(i), it is shifted back to the interval
0 
coords(i) < dims(i) automatically. Out-of-range
coordinates are erroneous for non-periodic dimensions.
MPI_CART_COORDS(comm, rank, maxdims, coords) IN comm communicator with Cartesian structure (handle) IN rank rank of a process within group of comm (integer) IN maxdims length of vector coords in the calling program (integer) OUT coords integer array (of size ndims) containing the Cartesian coordinates of specified process (array of integers)
INTEGER COMM, RANK, MAXDIMS, COORDS(*), IERROR
{ void MPI::Cartcomm::Get_coords(int rank, int maxdims, int coords[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
The inverse mapping, rank-to-coordinates translation is provided by
MPI_CART_COORDS.
If comm is associated with a zero-dimensional Cartesian topology,
coords will be unchanged.
MPI_GRAPH_NEIGHBORS_COUNT(comm, rank, nneighbors) IN comm communicator with graph topology (handle) IN rank rank of process in group of comm (integer) OUT nneighbors number of neighbors of specified process (integer)
INTEGER COMM, RANK, NNEIGHBORS, IERROR
{ int MPI::Graphcomm::Get_neighbors_count(int rank) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
2.2
MPI_GRAPH_NEIGHBORS(comm, rank, maxneighbors, neighbors) IN comm communicator with graph topology (handle) IN rank rank of process in group of comm (integer) IN maxneighbors size of array neighbors (integer) OUT neighbors ranks of processes that are neighbors to specified process (array of integer)
INTEGER COMM, RANK, MAXNEIGHBORS, NEIGHBORS(*), IERROR
{ void MPI::Graphcomm::Get_neighbors(int rank, int maxneighbors, int neighbors[]) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
MPI_GRAPH_NEIGHBORS_COUNT and MPI_GRAPH_NEIGHBORS provide
adjacency information for a general graph topology.
The returned count and array of neighbors for the queried rank will
both include all neighbors and reflect the same edge ordering as
was specified by the original call to MPI_GRAPH_CREATE.
Specifically, MPI_GRAPH_NEIGHBORS_COUNT and
MPI_GRAPH_NEIGHBORS will return values based on the original
index and edges array passed to MPI_GRAPH_CREATE
(assuming that index[-1] effectively equals zero):
Example
Assume there are four processes 0, 1, 2, 3 with the following
adjacency matrix (note that some neighbors are listed multiple times):
process neighbors
0 1, 1, 3
1 0, 0
2 3
3 0, 2, 2
nnodes = 4
index = 3, 5, 6, 9
edges = 1, 1, 3, 0, 0, 3, 0, 2, 2
Input rank Count
Neighbors
0 3 1, 1, 3
1 2 0, 0
2 1 3
3 3 0, 2, 2
Example
Suppose that comm is a communicator with a
shuffle-exchange topology. The group has 2n members.
Each process is labeled by a1 , ..., an with 
, and has three neighbors:
exchange(
(
), shuffle(a1 , ..., an )= a2 , ...,
an, a1, and unshuffle(a1 , ..., an ) = an , a1 , ... , an-1.
The graph adjacency list is illustrated below for n=3.
| node | exchange | shuffle | unshuffle | ||
| neighbors(1) | neighbors(2) | neighbors(3) | |||
| 0 | (000) | 1 | 0 | 0 | |
| 1 | (001) | 0 | 2 | 4 | |
| 2 | (010) | 3 | 4 | 1 | |
| 3 | (011) | 2 | 6 | 5 | |
| 4 | (100) | 5 | 1 | 2 | |
| 5 | (101) | 4 | 3 | 6 | |
| 6 | (110) | 7 | 5 | 3 | |
| 7 | (111) | 6 | 7 | 7 | |
Suppose that the communicator comm has this topology associated with it. The following code fragment cycles through the three types of neighbors and performs an appropriate permutation for each.
C assume: each process has stored a real number A.
C extract neighborhood information
CALL MPI_COMM_RANK(comm, myrank, ierr)
CALL MPI_GRAPH_NEIGHBORS(comm, myrank, 3, neighbors, ierr)
C perform exchange permutation
CALL MPI_SENDRECV_REPLACE(A, 1, MPI_REAL, neighbors(1), 0,
+ neighbors(1), 0, comm, status, ierr)
C perform shuffle permutation
CALL MPI_SENDRECV_REPLACE(A, 1, MPI_REAL, neighbors(2), 0,
+ neighbors(3), 0, comm, status, ierr)
C perform unshuffle permutation
CALL MPI_SENDRECV_REPLACE(A, 1, MPI_REAL, neighbors(3), 0,
+ neighbors(2), 0, comm, status, ierr)
MPI_DIST_GRAPH_NEIGHBORS_COUNT and MPI_DIST_GRAPH_NEIGHBORS
provide adjacency information for a distributed graph topology.
int MPI_Dist_graph_neighbors_count(MPI_Comm comm, int *indegree, int *outdegree, int *weighted)
MPI_DIST_GRAPH_NEIGHBORS_COUNT(COMM, INDEGREE, OUTDEGREE, WEIGHTED, IERROR)
{ void MPI::Distgraphcomm::Get_dist_neighbors_count(int rank, int indegree[], int outdegree[], bool& weighted) const (binding deprecated, see Section Deprecated since MPI-2.2
) }
int MPI_Dist_graph_neighbors(MPI_Comm comm, int maxindegree, int sources[], int sourceweights[], int maxoutdegree, int destinations[], int destweights[])
MPI_DIST_GRAPH_NEIGHBORS(COMM, MAXINDEGREE, SOURCES, SOURCEWEIGHTS, MAXOUTDEGREE, DESTINATIONS, DESTWEIGHTS, IERROR)
{ void MPI::Distgraphcomm::Get_dist_neighbors(int maxindegree, int sources[], int sourceweights[], int maxoutdegree, int destinations[], int destweights[]) (binding deprecated, see Section Deprecated since MPI-2.2
) }
These calls are local. The number of edges into and out of the process
returned by MPI_DIST_GRAPH_NEIGHBORS_COUNT are the total
number of such edges given in the call to
MPI_DIST_GRAPH_CREATE_ADJACENT or
MPI_DIST_GRAPH_CREATE (potentially by processes other
than the calling process in the case of
MPI_DIST_GRAPH_CREATE). Multiply defined edges are all
counted and returned by MPI_DIST_GRAPH_NEIGHBORS in some
order. If MPI_UNWEIGHTED is supplied for
sourceweights or destweights or both, or if
MPI_UNWEIGHTED was supplied during the construction of the
graph then no weight information is returned in that array or those
arrays. The only requirement on the order of values in
sources and destinations is that two calls to the
routine with same input argument comm will return the same
sequence of edges. If maxindegree or maxoutdegree is
smaller than the numbers returned by
MPI_DIST_GRAPH_NEIGHBOR_COUNT, then only the first part
of the full list is returned. Note, that the order of returned edges
does need not to be identical to the order that was provided in the
creation of comm for the case that
MPI_DIST_GRAPH_CREATE_ADJACENT was used.
Since the query calls are defined to be local, each process needs to
store the list of its neighbors with incoming and outgoing
edges. Communication is required at the collective
MPI_DIST_GRAPH_CREATE call in order to compute the neighbor
lists for each process from the distributed graph specification.
( End of advice to implementors.)
MPI_DIST_GRAPH_NEIGHBORS_COUNT(comm, indegree, outdegree, weighted) IN comm communicator with distributed graph topology (handle) OUT indegree number of edges into this process (non-negative integer) OUT outdegree number of edges out of this process (non-negative integer) OUT weighted false if MPI_UNWEIGHTED was supplied during creation,
true otherwise (logical)
INTEGER COMM, INDEGREE, OUTDEGREE, IERROR
LOGICAL WEIGHTED
MPI_DIST_GRAPH_NEIGHBORS(comm, maxindegree, sources, sourceweights,
maxoutdegree, destinations, destweights) IN comm communicator with distributed graph topology (handle) IN maxindegree size of sources and sourceweights arrays (non-negative integer) OUT sources processes for which the calling process is a
destination (array of non-negative integers) OUT sourceweights weights of the edges into the calling process
(array of non-negative integers) IN maxoutdegree size of destinations and destweights arrays (non-negative integer) OUT destinations processes for which the calling process is a source
(array of non-negative integers) OUT destweights weights of the edges out of the calling process
(array of non-negative integers)
INTEGER COMM, MAXINDEGREE, SOURCES(*), SOURCEWEIGHTS(*), MAXOUTDEGREE,
DESTINATIONS(*), DESTWEIGHTS(*), IERROR
Advice
to implementors.
Up: Topology Constructors
Next: Cartesian Shift Coordinates
Previous: Distributed (Graph) Constructor
Return to MPI-2.2 Standard Index
Return to MPI Forum Home Page
(Unofficial) MPI-2.2 of September 4, 2009
HTML Generated on September 10, 2009