Gathers data from and scatters data to all members of a group. The MPI_Alltoallv function adds flexibility to the MPI_Alltoall function by allowing a varying count of data from each process.
int MPIAPI MPI_Alltoallv( _In_ void *sendbuf, _In_ int *sendcounts, _In_ int *sdispls, MPI_Datatype sendtype, _Out_ void *recvbuf, _In_ int *recvcounts, _In_ int *rdispls, MPI_Datatype recvtype, MPI_Comm comm );
The pointer to the data to be sent to all processes in the group. The number and data type of the elements in the buffer are specified in the sendcount and sendtype parameters. Each element in the buffer corresponds to a process in the group.
If the comm parameter references an intracommunicator, you can specify an in-place option by specifying MPI_IN_PLACE in all processes. The sendcount, sdispls and sendtype parameters are ignored. Each process enters data in the corresponding receive buffer element.
Data that are sent and received must have the same type map as specified by the recvcounts array and the recvtype parameter. Data is read from the locations of the receive buffer that is specified by the rdispls parameter.
The number of data elements that this process sends in the buffer that is specified in the sendbuf parameter. If an element insendcount is zero, the data part of the message from that process is empty.
The location, relative to the sendbuf parameter, of the data for each communicator process.
Entry j specifies the displacement relative to the sendbuf parameter from which to take the outgoing data that is destined for process j.
The MPI data type of the elements in the send buffer.
The pointer to a buffer that contains the data that are received from each process. The number and data type of the elements in the buffer are specified in the recvcount and recvtype parameters.
The number of data elements from each communicator process in the receive buffer.
The location, relative to the recvbuf parameter, of the data from each communicator process.
In the recvbuf, recvcounts, and rdispls parameter arrays, the nth element of each array refers to the data that is received from the nth communicator process.
The MPI data type of each element in buffer.
Specifies the MPI_Comm communicator handle.
Returns MPI_SUCCESS on success. Otherwise, the return value is an error code.
In Fortran, the return value is stored in the IERROR parameter.
MPI_ALLTOALL(SENDBUF, SENDCOUNT, SENDTYPE, RECVBUF, RECVCOUNT, RECVTYPE, COMM, IERROR) <type> SENDBUF(*), R.ECVBUF(*) INTEGER SENDCOUNT, SENDTYPE, RECVCOUNT, RECVTYPE, COMM, IERROR
The type signature that is specified by the sendcount, and sendtype parameters for a process must be equal to the type signature that is specified by the recvcount, and recvtype parameters of the receiving process. Therefore, the amount of data sent must be equal to the amount of data that is received between any pair of processes. Distinct type maps between sender and receiver are still allowed.
All parameters are significant on all processes. The comm parameter must be identical on all processes.
If the comm parameter references an intracommunicator, then the jth block that is sent from process i is received by process j and is placed in the ith block of the receive buffer. These blocks do not have to all have the same size.
The outcome of calling the MPI_Alltoallv function is as if each process sent a message to every other process with,
MPI_Send(sendbuf + sdispls[i]*extent(sendtype), sendcounts[i], sendtype, I, …), and received a message from every other process with a call to
MPI_Recv(recvbuf + rdispls[i]*extent(recvtype), recvcounts[i], recvtype, I, …).
Specifying the in-place option indicates that the same amount and type of data is sent and received between any two processes in the group of the communicator. Different pairs of processes can exchange different amounts of data. Users must ensure that recvcounts[j] and recvtype on process i match recvcounts[i] and recvtype on process j. This symmetric exchange can be useful in applications where the data to be sent is not be used by the sending process after the MPI_Alltoallv function call.
If the comm parameter references an intercommunicator, then the outcome is as if each process in group A sends a message to each process in group B, and vice versa. The jth send buffer of process i in group A should be consistent with the ith receive buffer of process j in group B, and vice versa.
HPC Pack 2012 MS-MPI Redistributable Package, HPC Pack 2008 R2 MS-MPI Redistributable Package, HPC Pack 2008 MS-MPI Redistributable Package or HPC Pack 2008 Client Utilities