1
0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 15:55:37 +00:00
Grid/lib/lattice/Lattice_reduction.h
Azusa Yamaguchi d9cd4f0273 Staggered multinode block cg debugged. Missing global sum.
Code stalls and resumes on KNL at cambridge. Curious.

CG iterations 23ms each, then 3200 ms pauses. Mean bandwidth reports
as 200MB/s. Comms dominant in the report. However, the time behaviour suggests it
is *bursty*.... Could be swap to disk?
2017-08-23 15:07:18 +01:00

573 lines
16 KiB
C++

/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/lattice/Lattice_reduction.h
Copyright (C) 2015
Author: Azusa Yamaguchi <ayamaguc@staffmail.ed.ac.uk>
Author: Peter Boyle <paboyle@ph.ed.ac.uk>
Author: paboyle <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_LATTICE_REDUCTION_H
#define GRID_LATTICE_REDUCTION_H
#include <Grid/Grid_Eigen_Dense.h>
namespace Grid {
#ifdef GRID_WARN_SUBOPTIMAL
#warning "Optimisation alert all these reduction loops are NOT threaded "
#endif
////////////////////////////////////////////////////////////////////////////////////////////////////
// Deterministic Reduction operations
////////////////////////////////////////////////////////////////////////////////////////////////////
template<class vobj> inline RealD norm2(const Lattice<vobj> &arg){
ComplexD nrm = innerProduct(arg,arg);
return std::real(nrm);
}
// Double inner product
template<class vobj>
inline ComplexD innerProduct(const Lattice<vobj> &left,const Lattice<vobj> &right)
{
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_typeD vector_type;
scalar_type nrm;
GridBase *grid = left._grid;
std::vector<vector_type,alignedAllocator<vector_type> > sumarray(grid->SumArraySize());
parallel_for(int thr=0;thr<grid->SumArraySize();thr++){
int nwork, mywork, myoff;
GridThread::GetWork(left._grid->oSites(),thr,mywork,myoff);
decltype(innerProductD(left._odata[0],right._odata[0])) vnrm=zero; // private to thread; sub summation
for(int ss=myoff;ss<mywork+myoff; ss++){
vnrm = vnrm + innerProductD(left._odata[ss],right._odata[ss]);
}
sumarray[thr]=TensorRemove(vnrm) ;
}
vector_type vvnrm; vvnrm=zero; // sum across threads
for(int i=0;i<grid->SumArraySize();i++){
vvnrm = vvnrm+sumarray[i];
}
nrm = Reduce(vvnrm);// sum across simd
right._grid->GlobalSum(nrm);
return nrm;
}
template<class Op,class T1>
inline auto sum(const LatticeUnaryExpression<Op,T1> & expr)
->typename decltype(expr.first.func(eval(0,std::get<0>(expr.second))))::scalar_object
{
return sum(closure(expr));
}
template<class Op,class T1,class T2>
inline auto sum(const LatticeBinaryExpression<Op,T1,T2> & expr)
->typename decltype(expr.first.func(eval(0,std::get<0>(expr.second)),eval(0,std::get<1>(expr.second))))::scalar_object
{
return sum(closure(expr));
}
template<class Op,class T1,class T2,class T3>
inline auto sum(const LatticeTrinaryExpression<Op,T1,T2,T3> & expr)
->typename decltype(expr.first.func(eval(0,std::get<0>(expr.second)),
eval(0,std::get<1>(expr.second)),
eval(0,std::get<2>(expr.second))
))::scalar_object
{
return sum(closure(expr));
}
template<class vobj>
inline typename vobj::scalar_object sum(const Lattice<vobj> &arg)
{
GridBase *grid=arg._grid;
int Nsimd = grid->Nsimd();
std::vector<vobj,alignedAllocator<vobj> > sumarray(grid->SumArraySize());
for(int i=0;i<grid->SumArraySize();i++){
sumarray[i]=zero;
}
parallel_for(int thr=0;thr<grid->SumArraySize();thr++){
int nwork, mywork, myoff;
GridThread::GetWork(grid->oSites(),thr,mywork,myoff);
vobj vvsum=zero;
for(int ss=myoff;ss<mywork+myoff; ss++){
vvsum = vvsum + arg._odata[ss];
}
sumarray[thr]=vvsum;
}
vobj vsum=zero; // sum across threads
for(int i=0;i<grid->SumArraySize();i++){
vsum = vsum+sumarray[i];
}
typedef typename vobj::scalar_object sobj;
sobj ssum=zero;
std::vector<sobj> buf(Nsimd);
extract(vsum,buf);
for(int i=0;i<Nsimd;i++) ssum = ssum + buf[i];
arg._grid->GlobalSum(ssum);
return ssum;
}
//////////////////////////////////////////////////////////////////////////////////////////////////////////////
// sliceSum, sliceInnerProduct, sliceAxpy, sliceNorm etc...
//////////////////////////////////////////////////////////////////////////////////////////////////////////////
template<class vobj> inline void sliceSum(const Lattice<vobj> &Data,std::vector<typename vobj::scalar_object> &result,int orthogdim)
{
///////////////////////////////////////////////////////
// FIXME precision promoted summation
// may be important for correlation functions
// But easily avoided by using double precision fields
///////////////////////////////////////////////////////
typedef typename vobj::scalar_object sobj;
GridBase *grid = Data._grid;
assert(grid!=NULL);
const int Nd = grid->_ndimension;
const int Nsimd = grid->Nsimd();
assert(orthogdim >= 0);
assert(orthogdim < Nd);
int fd=grid->_fdimensions[orthogdim];
int ld=grid->_ldimensions[orthogdim];
int rd=grid->_rdimensions[orthogdim];
std::vector<vobj,alignedAllocator<vobj> > lvSum(rd); // will locally sum vectors first
std::vector<sobj> lsSum(ld,zero); // sum across these down to scalars
std::vector<sobj> extracted(Nsimd); // splitting the SIMD
result.resize(fd); // And then global sum to return the same vector to every node
for(int r=0;r<rd;r++){
lvSum[r]=zero;
}
int e1= grid->_slice_nblock[orthogdim];
int e2= grid->_slice_block [orthogdim];
int stride=grid->_slice_stride[orthogdim];
// sum over reduced dimension planes, breaking out orthog dir
// Parallel over orthog direction
parallel_for(int r=0;r<rd;r++){
int so=r*grid->_ostride[orthogdim]; // base offset for start of plane
for(int n=0;n<e1;n++){
for(int b=0;b<e2;b++){
int ss= so+n*stride+b;
lvSum[r]=lvSum[r]+Data._odata[ss];
}
}
}
// Sum across simd lanes in the plane, breaking out orthog dir.
std::vector<int> icoor(Nd);
for(int rt=0;rt<rd;rt++){
extract(lvSum[rt],extracted);
for(int idx=0;idx<Nsimd;idx++){
grid->iCoorFromIindex(icoor,idx);
int ldx =rt+icoor[orthogdim]*rd;
lsSum[ldx]=lsSum[ldx]+extracted[idx];
}
}
// sum over nodes.
sobj gsum;
for(int t=0;t<fd;t++){
int pt = t/ld; // processor plane
int lt = t%ld;
if ( pt == grid->_processor_coor[orthogdim] ) {
gsum=lsSum[lt];
} else {
gsum=zero;
}
grid->GlobalSum(gsum);
result[t]=gsum;
}
}
template<class vobj>
static void sliceInnerProductVector( std::vector<ComplexD> & result, const Lattice<vobj> &lhs,const Lattice<vobj> &rhs,int orthogdim)
{
typedef typename vobj::vector_type vector_type;
typedef typename vobj::scalar_type scalar_type;
GridBase *grid = lhs._grid;
assert(grid!=NULL);
conformable(grid,rhs._grid);
const int Nd = grid->_ndimension;
const int Nsimd = grid->Nsimd();
assert(orthogdim >= 0);
assert(orthogdim < Nd);
int fd=grid->_fdimensions[orthogdim];
int ld=grid->_ldimensions[orthogdim];
int rd=grid->_rdimensions[orthogdim];
std::vector<vector_type,alignedAllocator<vector_type> > lvSum(rd); // will locally sum vectors first
std::vector<scalar_type > lsSum(ld,scalar_type(0.0)); // sum across these down to scalars
std::vector<iScalar<scalar_type> > extracted(Nsimd); // splitting the SIMD
result.resize(fd); // And then global sum to return the same vector to every node for IO to file
for(int r=0;r<rd;r++){
lvSum[r]=zero;
}
int e1= grid->_slice_nblock[orthogdim];
int e2= grid->_slice_block [orthogdim];
int stride=grid->_slice_stride[orthogdim];
parallel_for(int r=0;r<rd;r++){
int so=r*grid->_ostride[orthogdim]; // base offset for start of plane
for(int n=0;n<e1;n++){
for(int b=0;b<e2;b++){
int ss= so+n*stride+b;
vector_type vv = TensorRemove(innerProduct(lhs._odata[ss],rhs._odata[ss]));
lvSum[r]=lvSum[r]+vv;
}
}
}
// Sum across simd lanes in the plane, breaking out orthog dir.
std::vector<int> icoor(Nd);
for(int rt=0;rt<rd;rt++){
iScalar<vector_type> temp;
temp._internal = lvSum[rt];
extract(temp,extracted);
for(int idx=0;idx<Nsimd;idx++){
grid->iCoorFromIindex(icoor,idx);
int ldx =rt+icoor[orthogdim]*rd;
lsSum[ldx]=lsSum[ldx]+extracted[idx]._internal;
}
}
// sum over nodes.
scalar_type gsum;
for(int t=0;t<fd;t++){
int pt = t/ld; // processor plane
int lt = t%ld;
if ( pt == grid->_processor_coor[orthogdim] ) {
gsum=lsSum[lt];
} else {
gsum=scalar_type(0.0);
}
grid->GlobalSum(gsum);
result[t]=gsum;
}
}
template<class vobj>
static void sliceNorm (std::vector<RealD> &sn,const Lattice<vobj> &rhs,int Orthog)
{
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
int Nblock = rhs._grid->GlobalDimensions()[Orthog];
std::vector<ComplexD> ip(Nblock);
sn.resize(Nblock);
sliceInnerProductVector(ip,rhs,rhs,Orthog);
for(int ss=0;ss<Nblock;ss++){
sn[ss] = real(ip[ss]);
}
};
template<class vobj>
static void sliceMaddVector(Lattice<vobj> &R,std::vector<RealD> &a,const Lattice<vobj> &X,const Lattice<vobj> &Y,
int orthogdim,RealD scale=1.0)
{
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
typedef typename vobj::tensor_reduced tensor_reduced;
scalar_type zscale(scale);
GridBase *grid = X._grid;
int Nsimd =grid->Nsimd();
int Nblock =grid->GlobalDimensions()[orthogdim];
int fd =grid->_fdimensions[orthogdim];
int ld =grid->_ldimensions[orthogdim];
int rd =grid->_rdimensions[orthogdim];
int e1 =grid->_slice_nblock[orthogdim];
int e2 =grid->_slice_block [orthogdim];
int stride =grid->_slice_stride[orthogdim];
std::vector<int> icoor;
for(int r=0;r<rd;r++){
int so=r*grid->_ostride[orthogdim]; // base offset for start of plane
vector_type av;
for(int l=0;l<Nsimd;l++){
grid->iCoorFromIindex(icoor,l);
int ldx =r+icoor[orthogdim]*rd;
scalar_type *as =(scalar_type *)&av;
as[l] = scalar_type(a[ldx])*zscale;
}
tensor_reduced at; at=av;
parallel_for_nest2(int n=0;n<e1;n++){
for(int b=0;b<e2;b++){
int ss= so+n*stride+b;
R._odata[ss] = at*X._odata[ss]+Y._odata[ss];
}
}
}
};
/*
inline GridBase *makeSubSliceGrid(const GridBase *BlockSolverGrid,int Orthog)
{
int NN = BlockSolverGrid->_ndimension;
int nsimd = BlockSolverGrid->Nsimd();
std::vector<int> latt_phys(0);
std::vector<int> simd_phys(0);
std::vector<int> mpi_phys(0);
for(int d=0;d<NN;d++){
if( d!=Orthog ) {
latt_phys.push_back(BlockSolverGrid->_fdimensions[d]);
simd_phys.push_back(BlockSolverGrid->_simd_layout[d]);
mpi_phys.push_back(BlockSolverGrid->_processors[d]);
}
}
return (GridBase *)new GridCartesian(latt_phys,simd_phys,mpi_phys);
}
*/
template<class vobj>
static void sliceMaddMatrix (Lattice<vobj> &R,Eigen::MatrixXcd &aa,const Lattice<vobj> &X,const Lattice<vobj> &Y,int Orthog,RealD scale=1.0)
{
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
int Nblock = X._grid->GlobalDimensions()[Orthog];
GridBase *FullGrid = X._grid;
// GridBase *SliceGrid = makeSubSliceGrid(FullGrid,Orthog);
// Lattice<vobj> Xslice(SliceGrid);
// Lattice<vobj> Rslice(SliceGrid);
assert( FullGrid->_simd_layout[Orthog]==1);
int nh = FullGrid->_ndimension;
// int nl = SliceGrid->_ndimension;
int nl = nh-1;
//FIXME package in a convenient iterator
//Should loop over a plane orthogonal to direction "Orthog"
int stride=FullGrid->_slice_stride[Orthog];
int block =FullGrid->_slice_block [Orthog];
int nblock=FullGrid->_slice_nblock[Orthog];
int ostride=FullGrid->_ostride[Orthog];
#pragma omp parallel
{
std::vector<vobj> s_x(Nblock);
#pragma omp for collapse(2)
for(int n=0;n<nblock;n++){
for(int b=0;b<block;b++){
int o = n*stride + b;
for(int i=0;i<Nblock;i++){
s_x[i] = X[o+i*ostride];
}
vobj dot;
for(int i=0;i<Nblock;i++){
dot = Y[o+i*ostride];
for(int j=0;j<Nblock;j++){
dot = dot + s_x[j]*(scale*aa(j,i));
}
R[o+i*ostride]=dot;
}
}}
}
};
template<class vobj>
static void sliceMulMatrix (Lattice<vobj> &R,Eigen::MatrixXcd &aa,const Lattice<vobj> &X,int Orthog,RealD scale=1.0)
{
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
int Nblock = X._grid->GlobalDimensions()[Orthog];
GridBase *FullGrid = X._grid;
// GridBase *SliceGrid = makeSubSliceGrid(FullGrid,Orthog);
// Lattice<vobj> Xslice(SliceGrid);
// Lattice<vobj> Rslice(SliceGrid);
assert( FullGrid->_simd_layout[Orthog]==1);
int nh = FullGrid->_ndimension;
// int nl = SliceGrid->_ndimension;
int nl=1;
//FIXME package in a convenient iterator
//Should loop over a plane orthogonal to direction "Orthog"
int stride=FullGrid->_slice_stride[Orthog];
int block =FullGrid->_slice_block [Orthog];
int nblock=FullGrid->_slice_nblock[Orthog];
int ostride=FullGrid->_ostride[Orthog];
#pragma omp parallel
{
std::vector<vobj> s_x(Nblock);
#pragma omp for collapse(2)
for(int n=0;n<nblock;n++){
for(int b=0;b<block;b++){
int o = n*stride + b;
for(int i=0;i<Nblock;i++){
s_x[i] = X[o+i*ostride];
}
vobj dot;
for(int i=0;i<Nblock;i++){
dot = s_x[0]*(scale*aa(0,i));
for(int j=1;j<Nblock;j++){
dot = dot + s_x[j]*(scale*aa(j,i));
}
R[o+i*ostride]=dot;
}
}}
}
};
template<class vobj>
static void sliceInnerProductMatrix( Eigen::MatrixXcd &mat, const Lattice<vobj> &lhs,const Lattice<vobj> &rhs,int Orthog)
{
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_type scalar_type;
typedef typename vobj::vector_type vector_type;
GridBase *FullGrid = lhs._grid;
// GridBase *SliceGrid = makeSubSliceGrid(FullGrid,Orthog);
int Nblock = FullGrid->GlobalDimensions()[Orthog];
// Lattice<vobj> Lslice(SliceGrid);
// Lattice<vobj> Rslice(SliceGrid);
mat = Eigen::MatrixXcd::Zero(Nblock,Nblock);
assert( FullGrid->_simd_layout[Orthog]==1);
int nh = FullGrid->_ndimension;
// int nl = SliceGrid->_ndimension;
int nl = nh-1;
//FIXME package in a convenient iterator
//Should loop over a plane orthogonal to direction "Orthog"
int stride=FullGrid->_slice_stride[Orthog];
int block =FullGrid->_slice_block [Orthog];
int nblock=FullGrid->_slice_nblock[Orthog];
int ostride=FullGrid->_ostride[Orthog];
typedef typename vobj::vector_typeD vector_typeD;
#pragma omp parallel
{
std::vector<vobj> Left(Nblock);
std::vector<vobj> Right(Nblock);
Eigen::MatrixXcd mat_thread = Eigen::MatrixXcd::Zero(Nblock,Nblock);
#pragma omp for collapse(2)
for(int n=0;n<nblock;n++){
for(int b=0;b<block;b++){
int o = n*stride + b;
for(int i=0;i<Nblock;i++){
Left [i] = lhs[o+i*ostride];
Right[i] = rhs[o+i*ostride];
}
for(int i=0;i<Nblock;i++){
for(int j=0;j<Nblock;j++){
auto tmp = innerProduct(Left[i],Right[j]);
// vector_typeD rtmp = TensorRemove(tmp);
auto rtmp = TensorRemove(tmp);
mat_thread(i,j) += Reduce(rtmp);
}}
}}
#pragma omp critical
{
mat += mat_thread;
}
}
for(int i=0;i<Nblock;i++){
for(int j=0;j<Nblock;j++){
ComplexD sum = mat(i,j);
FullGrid->GlobalSum(sum);
mat(i,j)=sum;
}}
return;
}
} /*END NAMESPACE GRID*/
#endif