portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-17 15:27:06 +01:00
Code Releases Activity

Improvements to the assembler interface that let us move chunks of the

Browse Source 
site and s loop into the kernels. This will save on function call overhead and
guarantee L2 prefetching strategy is right since OMP can't distribute the
sub-chunks of work.
This commit is contained in:
paboyle
2016-06-09 01:12:36 -07:00
parent d9408893b3
commit 55f65b81b5
10 changed files with 77 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
lib/qcd/action/fermion/WilsonKernelsHand.cc
View File

@ -807,7 +807,7 @@ void WilsonKernels<GparityWilsonImplF>::DiracOptHandDhopSite(StencilImpl &st,Dou
std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> > &buf,
int sF,int sU,const FermionField &in, FermionField &out)
{
DiracOptDhopSite(st,U,buf,sF,sU,in,out); // returns void, will template override for Wilson Nc=3
assert(0);
}
template<>
@ -815,7 +815,7 @@ void WilsonKernels<GparityWilsonImplF>::DiracOptHandDhopSiteDag(StencilImpl &st,
std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> > &buf,
int sF,int sU,const FermionField &in, FermionField &out)
{
DiracOptDhopSiteDag(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
assert(0);
}
template<>
@ -823,7 +823,7 @@ void WilsonKernels<GparityWilsonImplD>::DiracOptHandDhopSite(StencilImpl &st,Dou
std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> > &buf,
int sF,int sU,const FermionField &in, FermionField &out)
{
DiracOptDhopSite(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
assert(0);
}
template<>
@ -831,7 +831,7 @@ void WilsonKernels<GparityWilsonImplD>::DiracOptHandDhopSiteDag(StencilImpl &st,
std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> > &buf,
int sF,int sU,const FermionField &in, FermionField &out)
{
DiracOptDhopSiteDag(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
assert(0);
}