cannam@95: Previous: FFTW MPI Reference, cannam@95: Up: Distributed-memory FFTW with MPI cannam@95:
cannam@95: The FFTW MPI interface is callable from modern Fortran compilers
cannam@95: supporting the Fortran 2003 iso_c_binding standard for calling
cannam@95: C functions. As described in Calling FFTW from Modern Fortran,
cannam@95: this means that you can directly call FFTW's C interface from Fortran
cannam@95: with only minor changes in syntax. There are, however, a few things
cannam@95: specific to the MPI interface to keep in mind:
cannam@95:
cannam@95:
fftw3.f03 as in Overview of Fortran interface, you should include 'fftw3-mpi.f03' (after
cannam@95: use, intrinsic :: iso_c_binding as before). The
cannam@95: fftw3-mpi.f03 file includes fftw3.f03, so you should
cannam@95: not include them both yourself. (You will also want to
cannam@95: include the MPI header file, usually via include 'mpif.h' or
cannam@95: similar, although though this is not needed by fftw3-mpi.f03
cannam@95: per se.) (To use the ‘fftwl_’ long double extended-precision routines in supporting compilers, you should include fftw3f-mpi.f03 in addition to fftw3-mpi.f03. See Extended and quadruple precision in Fortran.)
cannam@95:
cannam@95: integer types; there is
cannam@95: no MPI_Comm type, nor is there any way to access a C
cannam@95: MPI_Comm. Fortunately, this is taken care of for you by the
cannam@95: FFTW Fortran interface: whenever the C interface expects an
cannam@95: MPI_Comm type, you should pass the Fortran communicator as an
cannam@95: integer.1
cannam@95:
cannam@95: ptrdiff_t in C, you should use integer(C_INTPTR_T) in
cannam@95: Fortran (see FFTW Fortran type reference).
cannam@95:
cannam@95: fftw_execute_dft becomes fftw_mpi_execute_dft,
cannam@95: etcetera. See Using MPI Plans.
cannam@95:
cannam@95: For example, here is a Fortran code snippet to perform a distributed
cannam@95: L × M complex DFT in-place. (This assumes you have already
cannam@95: initialized MPI with MPI_init and have also performed
cannam@95: call fftw_mpi_init.)
cannam@95:
cannam@95:
use, intrinsic :: iso_c_binding cannam@95: include 'fftw3-mpi.f03' cannam@95: integer(C_INTPTR_T), parameter :: L = ... cannam@95: integer(C_INTPTR_T), parameter :: M = ... cannam@95: type(C_PTR) :: plan, cdata cannam@95: complex(C_DOUBLE_COMPLEX), pointer :: data(:,:) cannam@95: integer(C_INTPTR_T) :: i, j, alloc_local, local_M, local_j_offset cannam@95: cannam@95: ! get local data size and allocate (note dimension reversal) cannam@95: alloc_local = fftw_mpi_local_size_2d(M, L, MPI_COMM_WORLD, & cannam@95: local_M, local_j_offset) cannam@95: cdata = fftw_alloc_complex(alloc_local) cannam@95: call c_f_pointer(cdata, data, [L,local_M]) cannam@95: cannam@95: ! create MPI plan for in-place forward DFT (note dimension reversal) cannam@95: plan = fftw_mpi_plan_dft_2d(M, L, data, data, MPI_COMM_WORLD, & cannam@95: FFTW_FORWARD, FFTW_MEASURE) cannam@95: cannam@95: ! initialize data to some function my_function(i,j) cannam@95: do j = 1, local_M cannam@95: do i = 1, L cannam@95: data(i, j) = my_function(i, j + local_j_offset) cannam@95: end do cannam@95: end do cannam@95: cannam@95: ! compute transform (as many times as desired) cannam@95: call fftw_mpi_execute_dft(plan, data, data) cannam@95: cannam@95: call fftw_destroy_plan(plan) cannam@95: call fftw_free(cdata) cannam@95:cannam@95:
Note that when we called fftw_mpi_local_size_2d and
cannam@95: fftw_mpi_plan_dft_2d with the dimensions in reversed order,
cannam@95: since a L × M Fortran array is viewed by FFTW in C as a
cannam@95: M × L array. This means that the array was distributed over
cannam@95: the M dimension, the local portion of which is a
cannam@95: L × local_M array in Fortran. (You must not use an
cannam@95: allocate statement to allocate an L × local_M array,
cannam@95: however; you must allocate alloc_local complex numbers, which
cannam@95: may be greater than L * local_M, in order to reserve space for
cannam@95: intermediate steps of the transform.) Finally, we mention that
cannam@95: because C's array indices are zero-based, the local_j_offset
cannam@95: argument can conveniently be interpreted as an offset in the 1-based
cannam@95: j index (rather than as a starting index as in C).
cannam@95:
cannam@95:
If instead you had used the ior(FFTW_MEASURE,
cannam@95: FFTW_MPI_TRANSPOSED_OUT) flag, the output of the transform would be a
cannam@95: transposed M × local_L array, associated with the same
cannam@95: cdata allocation (since the transform is in-place), and which
cannam@95: you could declare with:
cannam@95:
cannam@95:
complex(C_DOUBLE_COMPLEX), pointer :: tdata(:,:) cannam@95: ... cannam@95: call c_f_pointer(cdata, tdata, [M,local_L]) cannam@95:cannam@95:
where local_L would have been obtained by changing the
cannam@95: fftw_mpi_local_size_2d call to:
cannam@95:
cannam@95:
alloc_local = fftw_mpi_local_size_2d_transposed(M, L, MPI_COMM_WORLD, & cannam@95: local_M, local_j_offset, local_L, local_i_offset) cannam@95:cannam@95:
[1] Technically, this is because you aren't
cannam@95: actually calling the C functions directly. You are calling wrapper
cannam@95: functions that translate the communicator with MPI_Comm_f2c
cannam@95: before calling the ordinary C interface. This is all done
cannam@95: transparently, however, since the fftw3-mpi.f03 interface file
cannam@95: renames the wrappers so that they are called in Fortran with the same
cannam@95: names as the C interface functions.