portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-16 14:57:05 +01:00
Code Releases Activity

Got unpreconditioned conjugate gradient to run and converge on a random (uniform random,

Browse Source 
not even SU(3) for now) gauge field. Convergence history is correctly indepdendent of decomposition
on 1,2,4,8,16 mpi tasks.
Found a couple of simd bugs which required fixed and enhanced the Grid_simd.cc test suite.
Implemented the Mdag, M, MdagM, Meooe Mooee schur type stuff in the wilson dop.
This commit is contained in:
Peter Boyle
2015-05-19 13:57:35 +01:00
parent 05d862782f
commit 4dba8522a1
33 changed files with 566 additions and 316 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
lib/lattice/Grid_lattice_arith.h
View File

@ -144,14 +144,44 @@ PARALLEL_FOR_LOOP
}
template<class sobj,class vobj> strong_inline
void axpy(Lattice<vobj> &ret,sobj a,const Lattice<vobj> &lhs,const Lattice<vobj> &rhs){
conformable(lhs,rhs);
void axpy(Lattice<vobj> &ret,sobj a,const Lattice<vobj> &x,const Lattice<vobj> &y){
conformable(x,y);
#pragma omp parallel for
for(int ss=0;ss<lhs._grid->oSites();ss++){
vobj tmp = a*lhs._odata[ss];
vstream(ret._odata[ss],tmp+rhs._odata[ss]);
for(int ss=0;ss<x._grid->oSites();ss++){
vobj tmp = a*x._odata[ss]+y._odata[ss];
vstream(ret._odata[ss],tmp);
}
}
template<class sobj,class vobj> strong_inline
void axpby(Lattice<vobj> &ret,sobj a,sobj b,const Lattice<vobj> &x,const Lattice<vobj> &y){
conformable(x,y);
#pragma omp parallel for
for(int ss=0;ss<x._grid->oSites();ss++){
vobj tmp = a*x._odata[ss]+b*y._odata[ss];
vstream(ret._odata[ss],tmp);
}
}
template<class sobj,class vobj> strong_inline
RealD axpy_norm(Lattice<vobj> &ret,sobj a,const Lattice<vobj> &x,const Lattice<vobj> &y){
conformable(x,y);
#pragma omp parallel for
for(int ss=0;ss<x._grid->oSites();ss++){
vobj tmp = a*x._odata[ss]+y._odata[ss];
vstream(ret._odata[ss],tmp);
}
return norm2(ret);
}
template<class sobj,class vobj> strong_inline
RealD axpby_norm(Lattice<vobj> &ret,sobj a,sobj b,const Lattice<vobj> &x,const Lattice<vobj> &y){
conformable(x,y);
#pragma omp parallel for
for(int ss=0;ss<x._grid->oSites();ss++){
vobj tmp = a*x._odata[ss]+b*y._odata[ss];
vstream(ret._odata[ss],tmp);
}
return norm2(ret); // FIXME implement parallel norm in ss loop
}
}
#endif