1
0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-15 02:05:37 +00:00
Grid/lib/algorithms/FFT.h

306 lines
8.7 KiB
C
Raw Normal View History

2016-08-17 01:33:55 +01:00
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/Cshift.h
Copyright (C) 2015
Author: Peter Boyle <paboyle@ph.ed.ac.uk>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
See the full license in the file "LICENSE" in the top level distribution directory
*************************************************************************************/
/* END LEGAL */
#ifndef _GRID_FFT_H_
#define _GRID_FFT_H_
2016-11-08 15:31:47 +00:00
#ifdef HAVE_FFTW
#ifdef USE_MKL
#include <fftw/fftw3.h>
#else
#include <fftw3.h>
#endif
2016-11-08 15:31:47 +00:00
#endif
2016-08-31 00:24:52 +01:00
2016-08-17 01:33:55 +01:00
namespace Grid {
template<class scalar> struct FFTW { };
#ifdef HAVE_FFTW
template<> struct FFTW<ComplexD> {
public:
typedef fftw_complex FFTW_scalar;
typedef fftw_plan FFTW_plan;
static FFTW_plan fftw_plan_many_dft(int rank, const int *n,int howmany,
FFTW_scalar *in, const int *inembed,
int istride, int idist,
FFTW_scalar *out, const int *onembed,
int ostride, int odist,
int sign, unsigned flags) {
return ::fftw_plan_many_dft(rank,n,howmany,in,inembed,istride,idist,out,onembed,ostride,odist,sign,flags);
}
static void fftw_flops(const FFTW_plan p,double *add, double *mul, double *fmas){
::fftw_flops(p,add,mul,fmas);
}
inline static void fftw_execute_dft(const FFTW_plan p,FFTW_scalar *in,FFTW_scalar *out) {
::fftw_execute_dft(p,in,out);
}
inline static void fftw_destroy_plan(const FFTW_plan p) {
::fftw_destroy_plan(p);
}
};
template<> struct FFTW<ComplexF> {
public:
typedef fftwf_complex FFTW_scalar;
typedef fftwf_plan FFTW_plan;
static FFTW_plan fftw_plan_many_dft(int rank, const int *n,int howmany,
FFTW_scalar *in, const int *inembed,
int istride, int idist,
FFTW_scalar *out, const int *onembed,
int ostride, int odist,
int sign, unsigned flags) {
return ::fftwf_plan_many_dft(rank,n,howmany,in,inembed,istride,idist,out,onembed,ostride,odist,sign,flags);
}
static void fftw_flops(const FFTW_plan p,double *add, double *mul, double *fmas){
::fftwf_flops(p,add,mul,fmas);
}
inline static void fftw_execute_dft(const FFTW_plan p,FFTW_scalar *in,FFTW_scalar *out) {
::fftwf_execute_dft(p,in,out);
}
inline static void fftw_destroy_plan(const FFTW_plan p) {
::fftwf_destroy_plan(p);
}
};
2016-08-24 16:41:47 +01:00
#endif
#ifndef FFTW_FORWARD
#define FFTW_FORWARD (-1)
#define FFTW_BACKWARD (+1)
#endif
2016-10-26 17:36:26 +01:00
class FFT {
2016-08-17 01:33:55 +01:00
private:
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
GridCartesian *vgrid;
GridCartesian *sgrid;
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
int Nd;
double flops;
double flops_call;
uint64_t usec;
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
std::vector<int> dimensions;
std::vector<int> processors;
std::vector<int> processor_coor;
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
public:
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
static const int forward=FFTW_FORWARD;
static const int backward=FFTW_BACKWARD;
2016-10-26 17:36:26 +01:00
double Flops(void) {return flops;}
double MFlops(void) {return flops/usec;}
2016-11-08 11:35:41 +00:00
double USec(void) {return (double)usec;}
2016-10-26 17:36:26 +01:00
FFT ( GridCartesian * grid ) :
vgrid(grid),
Nd(grid->_ndimension),
dimensions(grid->_fdimensions),
processors(grid->_processors),
processor_coor(grid->_processor_coor)
2016-08-17 01:33:55 +01:00
{
flops=0;
usec =0;
2016-08-17 01:33:55 +01:00
std::vector<int> layout(Nd,1);
sgrid = new GridCartesian(dimensions,layout,processors);
};
2016-10-26 17:36:26 +01:00
~FFT ( void) {
delete sgrid;
2016-08-17 01:33:55 +01:00
}
template<class vobj>
2016-08-31 00:24:52 +01:00
void FFT_dim_mask(Lattice<vobj> &result,const Lattice<vobj> &source,std::vector<int> mask,int sign){
conformable(result._grid,vgrid);
conformable(source._grid,vgrid);
Lattice<vobj> tmp(vgrid);
tmp = source;
for(int d=0;d<Nd;d++){
if( mask[d] ) {
FFT_dim(result,tmp,d,sign);
tmp=result;
}
}
}
template<class vobj>
void FFT_all_dim(Lattice<vobj> &result,const Lattice<vobj> &source,int sign){
std::vector<int> mask(Nd,1);
FFT_dim_mask(result,source,mask,sign);
}
template<class vobj>
void FFT_dim(Lattice<vobj> &result,const Lattice<vobj> &source,int dim, int sign){
2016-10-26 17:36:26 +01:00
#ifndef HAVE_FFTW
assert(0);
#else
2016-08-17 01:33:55 +01:00
conformable(result._grid,vgrid);
conformable(source._grid,vgrid);
int L = vgrid->_ldimensions[dim];
int G = vgrid->_fdimensions[dim];
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
std::vector<int> layout(Nd,1);
std::vector<int> pencil_gd(vgrid->_fdimensions);
2016-10-26 17:36:26 +01:00
pencil_gd[dim] = G*processors[dim];
2016-08-17 01:33:55 +01:00
// Pencil global vol LxLxGxLxL per node
GridCartesian pencil_g(pencil_gd,layout,processors);
2016-10-26 17:36:26 +01:00
2016-08-17 01:33:55 +01:00
// Construct pencils
typedef typename vobj::scalar_object sobj;
typedef typename sobj::scalar_type scalar;
2016-10-26 17:36:26 +01:00
Lattice<sobj> pgbuf(&pencil_g);
typedef typename FFTW<scalar>::FFTW_scalar FFTW_scalar;
typedef typename FFTW<scalar>::FFTW_plan FFTW_plan;
2016-10-26 17:36:26 +01:00
int Ncomp = sizeof(sobj)/sizeof(scalar);
int Nlow = 1;
for(int d=0;d<dim;d++){
Nlow*=vgrid->_ldimensions[d];
}
int rank = 1; /* 1d transforms */
int n[] = {G}; /* 1d transforms of length G */
int howmany = Ncomp;
int odist,idist,istride,ostride;
idist = odist = 1; /* Distance between consecutive FT's */
istride = ostride = Ncomp*Nlow; /* distance between two elements in the same FT */
int *inembed = n, *onembed = n;
scalar div;
if ( sign == backward ) div = 1.0/G;
else if ( sign == forward ) div = 1.0;
else assert(0);
2016-10-26 17:36:26 +01:00
FFTW_plan p;
{
2016-10-26 17:36:26 +01:00
FFTW_scalar *in = (FFTW_scalar *)&pgbuf._odata[0];
FFTW_scalar *out= (FFTW_scalar *)&pgbuf._odata[0];
p = FFTW<scalar>::fftw_plan_many_dft(rank,n,howmany,
in,inembed,
istride,idist,
out,onembed,
ostride, odist,
sign,FFTW_ESTIMATE);
}
// Barrel shift and collect global pencil
std::vector<int> lcoor(Nd), gcoor(Nd);
result = source;
for(int p=0;p<processors[dim];p++) {
2016-10-26 19:46:36 +01:00
PARALLEL_REGION
{
std::vector<int> cbuf(Nd);
2016-10-26 17:36:26 +01:00
sobj s;
2016-10-26 19:46:36 +01:00
PARALLEL_FOR_LOOP_INTERN
for(int idx=0;idx<sgrid->lSites();idx++) {
sgrid->LocalIndexToLocalCoor(idx,cbuf);
peekLocalSite(s,result,cbuf);
cbuf[dim]+=p*L;
pokeLocalSite(s,pgbuf,cbuf);
}
2016-10-26 17:36:26 +01:00
}
2016-11-24 00:09:48 +00:00
if (p != processors[dim] - 1)
{
result = Cshift(result,dim,L);
}
2016-10-26 17:36:26 +01:00
}
// Loop over orthog coords
int NN=pencil_g.lSites();
GridStopWatch timer;
timer.Start();
2016-10-26 19:46:36 +01:00
PARALLEL_REGION
{
std::vector<int> cbuf(Nd);
2016-10-26 17:36:26 +01:00
2016-10-26 19:46:36 +01:00
PARALLEL_FOR_LOOP_INTERN
for(int idx=0;idx<NN;idx++) {
pencil_g.LocalIndexToLocalCoor(idx, cbuf);
if ( cbuf[dim] == 0 ) { // restricts loop to plane at lcoor[dim]==0
FFTW_scalar *in = (FFTW_scalar *)&pgbuf._odata[idx];
FFTW_scalar *out= (FFTW_scalar *)&pgbuf._odata[idx];
FFTW<scalar>::fftw_execute_dft(p,in,out);
}
2016-10-26 17:36:26 +01:00
}
2016-08-17 01:33:55 +01:00
}
2016-10-26 17:36:26 +01:00
timer.Stop();
// performance counting
double add,mul,fma;
FFTW<scalar>::fftw_flops(p,&add,&mul,&fma);
flops_call = add+mul+2.0*fma;
usec += timer.useconds();
flops+= flops_call*NN;
// writing out result
int pc = processor_coor[dim];
2016-10-26 19:46:36 +01:00
PARALLEL_REGION
{
std::vector<int> clbuf(Nd), cgbuf(Nd);
2016-10-26 17:36:26 +01:00
sobj s;
2016-10-26 19:46:36 +01:00
PARALLEL_FOR_LOOP_INTERN
for(int idx=0;idx<sgrid->lSites();idx++) {
sgrid->LocalIndexToLocalCoor(idx,clbuf);
cgbuf = clbuf;
cgbuf[dim] = clbuf[dim]+L*pc;
peekLocalSite(s,pgbuf,cgbuf);
pokeLocalSite(s,result,clbuf);
}
2016-10-26 17:36:26 +01:00
}
2016-11-24 00:09:48 +00:00
result = result*div;
2016-10-26 17:36:26 +01:00
// destroying plan
FFTW<scalar>::fftw_destroy_plan(p);
#endif
2016-08-17 01:33:55 +01:00
}
};
}
#endif