pfft_oct_m Module Reference

The low level module to work with the PFFT library. http: More...

Detailed Description

The low level module to work with the PFFT library. http:

Functions/Subroutines
subroutine, public	pfft_decompose (n_proc, dim1, dim2)
	Decompose all available processors in 2D processor grid, most equally possible. More...
subroutine, public	pfft_prepare_plan_r2c (plan, n, in, out, sign, flags, mpi_comm)
	Octopus subroutine to prepare a PFFT plan real to complex. More...
subroutine, public	pfft_prepare_plan_c2r (plan, n, in, out, sign, flags, mpi_comm)
	Octopus subroutine to prepare a PFFT plan real to complex. More...
subroutine, public	pfft_prepare_plan_c2c (plan, n, in, out, sign, flags, mpi_comm)
	Octopus subroutine to prepare a PFFT plan real to complex. More...
subroutine, public	pfft_get_dims (rs_n_global, mpi_comm, is_real, alloc_size, fs_n_global, rs_n, fs_n, rs_istart, fs_istart)

Function/Subroutine Documentation

pfft_decompose()

subroutine, public pfft_oct_m::pfft_decompose	(	integer, intent(in)	n_proc,
		integer, intent(out)	dim1,
		integer, intent(out)	dim2
	)

Decompose all available processors in 2D processor grid, most equally possible.

Parameters

[in]	n_proc	Number of processors
[out]	dim1	First out dimension
[out]	dim2	Second out dimension

Definition at line 149 of file pfft.F90.

pfft_prepare_plan_r2c()

subroutine, public pfft_oct_m::pfft_prepare_plan_r2c	(	type(c_ptr), intent(out)	plan,
		integer, dimension(:), intent(in)	n,
		real(real64), dimension(:,:,:), intent(inout), pointer	in,
		complex(real64), dimension(:,:,:), intent(inout), pointer	out,
		integer, intent(in)	sign,
		integer, intent(in)	flags,
		integer, intent(in)	mpi_comm
	)

Octopus subroutine to prepare a PFFT plan real to complex.

Parameters

[out]	plan	The plan that is created by PFFT
[in]	n	The size of the global matrix
[in,out]	in	The input matrix that is going to be used to do the transform
[in,out]	out	The output matrix that is going to be used to do the transform
[in]	sign	Sign flag to decide FFT direction. Has to be FFTW_FORWARD
[in]	flags	Flags for FFT library. Could be changed with the input variable FFTPreparePlan. Default value is FFTW_MEASURE
[in]	mpi_comm	MPI communicator

Definition at line 190 of file pfft.F90.

pfft_prepare_plan_c2r()

subroutine, public pfft_oct_m::pfft_prepare_plan_c2r	(	type(c_ptr), intent(out)	plan,
		integer, dimension(:), intent(in)	n,
		complex(real64), dimension(:,:,:), intent(inout), pointer	in,
		real(real64), dimension(:,:,:), intent(inout), pointer	out,
		integer, intent(in)	sign,
		integer, intent(in)	flags,
		integer, intent(in)	mpi_comm
	)

Octopus subroutine to prepare a PFFT plan real to complex.

Parameters

[out]	plan	The plan that is created by PFFT
[in]	n	The size of the global matrix
[in,out]	in	The input matrix that is going to be used to do the transform
[in,out]	out	The output matrix that is going to be used to do the transform
[in]	sign	Sign flag to decide FFT direction. Has to be FFTW_BACKWARD
[in]	flags	Flags for FFT library. Could be changed with the input variable FFTPreparePlan. Default value is FFTW_MEASURE
[in]	mpi_comm	MPI communicator

Definition at line 224 of file pfft.F90.

pfft_prepare_plan_c2c()

subroutine, public pfft_oct_m::pfft_prepare_plan_c2c	(	type(c_ptr), intent(out)	plan,
		integer, dimension(:), intent(in)	n,
		complex(real64), dimension(:,:,:), intent(inout), pointer	in,
		complex(real64), dimension(:,:,:), intent(inout), pointer	out,
		integer, intent(in)	sign,
		integer, intent(in)	flags,
		integer, intent(in)	mpi_comm
	)

Octopus subroutine to prepare a PFFT plan real to complex.

Parameters

[out]	plan	The plan that is created by PFFT
[in]	n	The size of the global matrix
[in,out]	in	The input matrix that is going to be used to do the transform
[in,out]	out	The output matrix that is going to be used to do the transform
[in]	sign	Sign flag to decide FFT direction.
[in]	flags	Flags for FFT library. Could be changed with the input variable FFTPreparePlan. Default value is FFTW_MEASURE
[in]	mpi_comm	MPI communicator

Definition at line 259 of file pfft.F90.

pfft_get_dims()

subroutine, public pfft_oct_m::pfft_get_dims	(	integer, dimension(1:3), intent(in)	rs_n_global,
		integer, intent(in)	mpi_comm,
		logical, intent(in)	is_real,
		integer(int64), intent(out)	alloc_size,
		integer, dimension(1:3), intent(out)	fs_n_global,
		integer, dimension(1:3), intent(out)	rs_n,
		integer, dimension(1:3), intent(out)	fs_n,
		integer, dimension(1:3), intent(out)	rs_istart,
		integer, dimension(1:3), intent(out)	fs_istart
	)

Parameters

[in]	rs_n_global	The general number of elements in each dimension in real space
[in]	mpi_comm	MPI comunicator
[in]	is_real	The input is real or complex
[out]	alloc_size	Number of elements that has to be allocated locally
[out]	fs_n_global	The general number of elements in each dimension in Fourier space
[out]	rs_n	Local number of elements in each direction in real space
[out]	fs_n	Local number of elements in each direction in Fourier space
[out]	rs_istart	Where does the local portion of the function start in real space
[out]	fs_istart	Where does the local portion of the function start in Fourier space

Definition at line 286 of file pfft.F90.