portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-15 14:27:06 +01:00
Code Releases Activity

Threading support rework.

Browse Source 
Placed parallel pragmas as macros; implemented deterministic thread reduction in style of
BFM.
This commit is contained in:
Peter Boyle
2015-05-12 07:51:41 +01:00
parent b1d2c60d07
commit 6103c29ee3
26 changed files with 276 additions and 184 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
lib/lattice/Grid_lattice_base.h
View File

@ -64,7 +64,7 @@ public:
////////////////////////////////////////////////////////////////////////////////
template <typename Op, typename T1> inline Lattice<vobj> & operator=(const LatticeUnaryExpression<Op,T1> &expr)
{
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
vobj tmp= eval(ss,expr);
vstream(_odata[ss] ,tmp);
@ -73,7 +73,7 @@ public:
}
template <typename Op, typename T1,typename T2> inline Lattice<vobj> & operator=(const LatticeBinaryExpression<Op,T1,T2> &expr)
{
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
vobj tmp= eval(ss,expr);
vstream(_odata[ss] ,tmp);
@ -82,7 +82,7 @@ public:
}
template <typename Op, typename T1,typename T2,typename T3> inline Lattice<vobj> & operator=(const LatticeTrinaryExpression<Op,T1,T2,T3> &expr)
{
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
vobj tmp= eval(ss,expr);
vstream(_odata[ss] ,tmp);
@ -95,7 +95,7 @@ public:
GridFromExpression(_grid,expr);
assert(_grid!=nullptr);
_odata.resize(_grid->oSites());
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
_odata[ss] = eval(ss,expr);
}
@ -105,7 +105,7 @@ public:
GridFromExpression(_grid,expr);
assert(_grid!=nullptr);
_odata.resize(_grid->oSites());
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
_odata[ss] = eval(ss,expr);
}
@ -115,7 +115,7 @@ public:
GridFromExpression(_grid,expr);
assert(_grid!=nullptr);
_odata.resize(_grid->oSites());
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
_odata[ss] = eval(ss,expr);
}
@ -133,7 +133,7 @@ public:
}
template<class sobj> inline Lattice<vobj> & operator = (const sobj & r){
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
this->_odata[ss]=r;
}
@ -142,7 +142,7 @@ public:
template<class robj> inline Lattice<vobj> & operator = (const Lattice<robj> & r){
conformable(*this,r);
std::cout<<"Lattice operator ="<<std::endl;
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<_grid->oSites();ss++){
this->_odata[ss]=r._odata[ss];
}
@ -167,7 +167,7 @@ public:
inline friend Lattice<vobj> operator / (const Lattice<vobj> &lhs,const Lattice<vobj> &rhs){
conformable(lhs,rhs);
Lattice<vobj> ret(lhs._grid);
#pragma omp parallel for
PARALLEL_FOR_LOOP
for(int ss=0;ss<lhs._grid->oSites();ss++){
ret._odata[ss] = lhs._odata[ss]/rhs._odata[ss];
}