In applications like matrix transposes or Fast Fourier Transforms (FFTs), an MPI_Alltoall call is very helpful. This is an extension to MPI_Allgather where each process sends distinct data to each receiver. The j-th block from process i is received by process j and stored in the i-th block. A graphic representation of the MPI_Alltoall is shown below:

Just as with Allgatherv, the MPI_Alltoallv extension allows these blocks of data to vary in size among different pairs of senders and receivers. The syntax of MPI_Alltoall and MPI_Alltoallv is as follows:

Alltoall and Alltoallv Syntax

int MPI_Alltoall(void *sbuf, int scount, \
     MPI_Datatype stype, void *rbuf, int rcount, \
     MPI_Datatype rtype, MPI_Comm comm)

int MPI_Alltoallv(void *sbuf, int *scounts, \
     int *sdispls, MPI_Datatype stype, void *rbuf, \
     int *rcounts, int *rdispls, MPI_Datatype rtype, \
     MPI_Comm comm)

MPI_ALLTOALL(sbuf, scount, stype, rbuf, rcount, rtype, comm, ierr)

MPI_ALLTOALLV(sbuf, scounts, sdispls, stype, rbuf, rcounts,
     rdispls, rtype, comm, ierr)

Alltoall[v] parameters:

sbuf

is the starting address of the send buffer

scount[s]

is the [array of the] number of elements to send to each process

[sdispls]

is an array specifying the displacements of data relative to sbuf

stype

is the data type of send buffer elements

rbuf

is the address of the receive buffer

rcount[s]

is [an array containing] the number of elements to be received from each process

[rdispls]

is an array specifying the displacements of data relative to rbuf

rtype

is the data type of the receive buffer elements

comm

is the group communicator