portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-09-20 01:05:38 +01:00
Code Releases Activity

Merge pull request #365 from lehner/feature/gpt

Browse Source 
Sync
This commit is contained in:
Peter Boyle 2021-09-15 13:34:02 -04:00 committed by GitHub
commit 001a556a34
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
7 changed files with 148 additions and 136 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
Grid/algorithms/CoarsenedMatrix.h
View File

@ -442,6 +442,8 @@ public:
for(int p=0; p<geom.npoint; p++)
points[p] = geom.points_dagger[p];
auto points_p = &points[0];
RealD* dag_factor_p = &dag_factor[0];
accelerator_for(sss, Grid()->oSites()*nbasis, Nsimd, {
@ -453,7 +455,7 @@ public:
StencilEntry *SE;
for(int p=0;p<geom_v.npoint;p++){
int point = points[p];
int point = points_p[p];
SE=Stencil_v.GetEntry(ptype,point,ss);
@ -708,6 +710,8 @@ public:
for(int p=0; p<npoint; p++)
points[p] = (dag && !hermitian) ? geom.points_dagger[p] : p;
auto points_p = &points[0];
Vector<Aview> AcceleratorViewContainer;
for(int p=0;p<npoint;p++) AcceleratorViewContainer.push_back(a[p].View(AcceleratorRead));
Aview *Aview_p = & AcceleratorViewContainer[0];
@ -728,7 +732,7 @@ public:
StencilEntry *SE;
for(int p=0;p<npoint;p++){
int point = points[p];
int point = points_p[p];
SE=st_v.GetEntry(ptype,point,ss);
if(SE->_is_local) {
@ -754,7 +758,7 @@ public:
StencilEntry *SE;
for(int p=0;p<npoint;p++){
int point = points[p];
int point = points_p[p];
SE=st_v.GetEntry(ptype,point,ss);
if(SE->_is_local) {

4
Grid/algorithms/FFT.h
View File

@ -136,7 +136,7 @@ public:
flops=0;
usec =0;
Coordinate layout(Nd,1);
sgrid = new GridCartesian(dimensions,layout,processors);
sgrid = new GridCartesian(dimensions,layout,processors,*grid);
};
~FFT ( void) {
@ -182,7 +182,7 @@ public:
pencil_gd[dim] = G*processors[dim];
// Pencil global vol LxLxGxLxL per node
GridCartesian pencil_g(pencil_gd,layout,processors);
GridCartesian pencil_g(pencil_gd,layout,processors,*vgrid);
// Construct pencils
typedef typename vobj::scalar_object sobj;

2
Grid/lattice/Lattice_arith.h
View File

@ -225,7 +225,7 @@ void axpy(Lattice<vobj> &ret,sobj a,const Lattice<vobj> &x,const Lattice<vobj> &
autoView( x_v , x, AcceleratorRead);
autoView( y_v , y, AcceleratorRead);
accelerator_for(ss,x_v.size(),vobj::Nsimd(),{
auto tmp = a*x_v(ss)+y_v(ss);
auto tmp = a*coalescedRead(x_v[ss])+coalescedRead(y_v[ss]);
coalescedWrite(ret_v[ss],tmp);
});
}

4
Grid/lattice/Lattice_basis.h
View File

@ -125,7 +125,7 @@ void basisRotate(VField &basis,Matrix& Qt,int j0, int j1, int k0,int k1,int Nm)
for(int k=k0; k<k1; ++k){
auto tmp = coalescedRead(Bp[ss*nrot+j]);
coalescedWrite(Bp[ss*nrot+j],tmp+ Qt_p[jj*Nm+k] * coalescedRead(basis_v[k][sss]));
coalescedWrite(Bp[ss*nrot+j],tmp+ Qt_p[jj*Nm+k] * coalescedRead(basis_vp[k][sss]));
}
});
@ -134,7 +134,7 @@ void basisRotate(VField &basis,Matrix& Qt,int j0, int j1, int k0,int k1,int Nm)
int jj =j0+j;
int ss =sj/nrot;
int sss=ss+s;
coalescedWrite(basis_v[jj][sss],coalescedRead(Bp[ss*nrot+j]));
coalescedWrite(basis_vp[jj][sss],coalescedRead(Bp[ss*nrot+j]));
});
}
#endif

12
Grid/lattice/Lattice_reduction.h
View File

@ -361,6 +361,7 @@ template<class vobj> inline void sliceSum(const Lattice<vobj> &Data,std::vector<
// But easily avoided by using double precision fields
///////////////////////////////////////////////////////
typedef typename vobj::scalar_object sobj;
typedef typename vobj::scalar_object::scalar_type scalar_type;
GridBase *grid = Data.Grid();
assert(grid!=NULL);
@ -419,20 +420,19 @@ template<class vobj> inline void sliceSum(const Lattice<vobj> &Data,std::vector<
}
// 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];
result[t]=lsSum[lt];
} else {
gsum=Zero();
result[t]=Zero();
}
grid->GlobalSum(gsum);
result[t]=gsum;
}
scalar_type * ptr = (scalar_type *) &result[0];
int words = fd*sizeof(sobj)/sizeof(scalar_type);
grid->GlobalSumVector(ptr, words);
}
template<class vobj>

12
Grid/lattice/Lattice_transfer.h
View File

@ -364,16 +364,22 @@ inline void blockSum(Lattice<vobj> &coarseData,const Lattice<vobj> &fineData)
autoView( coarseData_ , coarseData, AcceleratorWrite);
autoView( fineData_ , fineData, AcceleratorRead);
auto coarseData_p = &coarseData_[0];
auto fineData_p = &fineData_[0];
Coordinate fine_rdimensions = fine->_rdimensions;
Coordinate coarse_rdimensions = coarse->_rdimensions;
vobj zz = Zero();
accelerator_for(sc,coarse->oSites(),1,{
// One thread per sub block
Coordinate coor_c(_ndimension);
Lexicographic::CoorFromIndex(coor_c,sc,coarse_rdimensions); // Block coordinate
coarseData_[sc]=Zero();
vobj cd = zz;
for(int sb=0;sb<blockVol;sb++){
int sf;
@ -383,9 +389,11 @@ inline void blockSum(Lattice<vobj> &coarseData,const Lattice<vobj> &fineData)
for(int d=0;d<_ndimension;d++) coor_f[d]=coor_c[d]*block_r[d] + coor_b[d];
Lexicographic::IndexFromCoor(coor_f,sf,fine_rdimensions);
coarseData_[sc]=coarseData_[sc]+fineData_[sf];
cd=cd+fineData_p[sf];
}
coarseData_p[sc] = cd;
});
return;
}

240
Grid/qcd/action/fermion/implementation/WilsonKernelsAsmA64FX.h
View File

@ -73,17 +73,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -102,17 +102,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -131,17 +131,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
@ -165,17 +165,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -194,17 +194,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFiel
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -223,17 +223,17 @@ WilsonKernels<ZWilsonImplF>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFiel
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplFH>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<WilsonImplFH>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
//#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
//template<> void
//WilsonKernels<ZWilsonImplFH>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
//#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
@ -280,17 +280,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSite(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -309,17 +309,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSiteInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -338,17 +338,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSiteExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
/////////////////////////////////////////////////////////////////
@ -371,17 +371,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldVi
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDag(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -400,17 +400,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFiel
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDagInt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#undef INTERIOR_AND_EXTERIOR
@ -429,17 +429,17 @@ WilsonKernels<ZWilsonImplD>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFiel
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<WilsonImplDF>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<WilsonImplDF>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
#pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
template<> void
WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
#include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>
// #pragma GCC optimize ("-O3", "-fno-schedule-insns", "-fno-schedule-insns2")
// template<> void
// WilsonKernels<ZWilsonImplDF>::AsmDhopSiteDagExt(StencilView &st, DoubledGaugeFieldView &U, SiteHalfSpinor *buf,
// int ss,int ssU,int Ls,int Ns,const FermionFieldView &in, FermionFieldView &out)
// #include <qcd/action/fermion/implementation/WilsonKernelsAsmBodyA64FX.h>