#include <mpi.h>
#include <omp.h>
#include <iostream>
#include <cmath>
#include <math.h>
#include <string.h>
#include <accfft.h>
#include <../src/operators.txx>
#include <pnetcdf.h>
#include <cstdlib>
#include <string>

Functions
void	accfft_grad (double A_x, double A_y, double A_z, double A, accfft_plan plan, std::bitset< 3 > pXYZ, double *timer)

void	accfft_laplace (double LA, double A, accfft_plan plan, double timer)

void	accfft_divergence (double divA, double A_x, double A_y, double A_z, accfft_plan plan, double timer)

void	accfft_biharmonic (double BA, double A, accfft_plan plan, double timer)

void	accfft_inv_laplace (double invLA, double A, accfft_plan plan, double timer)

void	accfft_inv_biharmonic (double invBA, double A, accfft_plan plan, double timer)

void	read_pnetcdf (const std::string &filename, MPI_Offset starts[3], MPI_Offset counts[3], int gsizes[3], double *localData)

void	write_pnetcdf (const std::string &filename, MPI_Offset starts[3], MPI_Offset counts[3], int gsizes[3], double *localData)

Detailed Description

CPU functions of AccFFT operators

Function Documentation

void accfft_biharmonic	(	double *	BA,
		double *	A,
		accfft_plan *	plan,
		double *	timer
	)

Computes double precision biharmonic of its input real data A, and writes the output into LA.

Parameters

BA	$\Delta^2 A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
timer	See Timing AccFFT for more details.

void accfft_divergence	(	double *	divA,
		double *	A_x,
		double *	A_y,
		double *	A_z,
		accfft_plan *	plan,
		double *	timer
	)

Computes double precision divergence of its input vector data A_x, A_y, and A_x. The output data is written to divA.

Parameters

divA	$\nabla\cdot(A_x i + A_y j+ A_z k)$
A_x	The x component of $\nabla A$
A_y	The y component of $\nabla A$
A_z	The z component of $\nabla A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
timer	See Timing AccFFT for more details.

void accfft_grad	(	double *	A_x,
		double *	A_y,
		double *	A_z,
		double *	A,
		accfft_plan *	plan,
		std::bitset< 3 > *	pXYZ,
		double *	timer
	)

Computes double precision gradient of its input real data A, and returns the x, y, and z components and writes the output into A_x, A_y, and A_z respectively.

Parameters

A_x	The x component of $\nabla A$
A_y	The y component of $\nabla A$
A_z	The z component of $\nabla A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
pXYZ	a bit set pointer field of size 3 that determines which gradient components are needed. If XYZ={111} then all the components are computed and if XYZ={100}, then only the x component is computed. This can save the user some time, when just one or two of the gradient components are needed.
timer	See Timing AccFFT for more details.

void accfft_inv_biharmonic	(	double *	invBA,
		double *	A,
		accfft_plan *	plan,
		double *	timer
	)

Computes double precision inverse biharmonic of its input real data A, and writes the output into invBA.

Parameters

invBA	$\Delta^{-2} A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
timer	See Timing AccFFT for more details.

void accfft_inv_laplace	(	double *	invLA,
		double *	A,
		accfft_plan *	plan,
		double *	timer
	)

Computes double precision inverse Laplacian of its input real data A, and writes the output into invLA.

Parameters

invLA	$\Delta^{-1} A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
timer	See Timing AccFFT for more details.

void accfft_laplace	(	double *	LA,
		double *	A,
		accfft_plan *	plan,
		double *	timer
	)

Computes double precision Laplacian of its input real data A, and writes the output into LA.

Parameters

LA	$\Delta A$
plan	FFT plan created by accfft_plan_dft_3d_r2c. Must be an outplace plan, otherwise the function will return without computing the gradient.
timer	See Timing AccFFT for more details.

void read_pnetcdf	(	const std::string &	filename,
		MPI_Offset	starts[3],
		MPI_Offset	counts[3],
		int	gsizes[3],
		double *	localData
	)

Read a parallel-nedcdf file.

We assume here that localData is a scalar.

Pnetcdf uses row-major format (same as FFTW).

Parameters

[in]	filename	: PnetCDF filename
[in]	starts	: offset to where to start reading data
[in]	counts	: number of elements read (3D sub-domain inside global)
[in]	gsizes	: global sizes
[out]	localData	: actual data buffer (size : nxnynz*sizeof(double))

localData must have been allocated prior to calling this routine.

void write_pnetcdf	(	const std::string &	filename,
		MPI_Offset	starts[3],
		MPI_Offset	counts[3],
		int	gsizes[3],
		double *	localData
	)

Write a parallel-nedcdf file.

We assume here that localData is a scalar.

Pnetcdf uses row-major format (same as FFTW).

Parameters

[in]	filename	: PnetCDF filename
[in]	starts	: offset to where to start reading data
[in]	counts	: number of elements read (3D sub-domain inside global)
[in]	gsizes	: global sizes
[in]	localData	: actual data buffer (size : nxnynz*sizeof(double))

Functions

Detailed Description

Function Documentation