portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-10-31 03:54:33 +00:00
Code Releases Activity

Thread loop changes

Browse Source
This commit is contained in:
Peter Boyle
2019-06-15 09:08:16 +01:00
parent 9a13d2992c
commit 52c74f1cac
1 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
Grid/qcd/action/fermion/implementation/ImprovedStaggeredFermionImplementation.h
View File

@@ -254,7 +254,7 @@ void ImprovedStaggeredFermion<Impl>::DerivInternal(StencilImpl &st, DoubledGauge
auto UUU_v = UUU.View(); auto UUU_v = UUU.View();
auto B_v = B.View(); auto B_v = B.View();
auto Btilde_v = Btilde.View(); auto Btilde_v = Btilde.View();
thread_loop( (int sss = 0; sss < B.Grid()->oSites(); sss++), { thread_for(sss,B.Grid()->oSites(),{
Kernels::DhopDirKernel(st, U_v, UUU_v, st.CommBuf(), sss, sss, B_v, Btilde_v, mu,1); Kernels::DhopDirKernel(st, U_v, UUU_v, st.CommBuf(), sss, sss, B_v, Btilde_v, mu,1);
}); });
@@ -375,7 +375,7 @@ void ImprovedStaggeredFermion<Impl>::DhopDir(const FermionField &in, FermionFiel
auto UUUmu_v = UUUmu.View(); auto UUUmu_v = UUUmu.View();
auto in_v = in.View(); auto in_v = in.View();
auto out_v = out.View(); auto out_v = out.View();
thread_loop( (int sss = 0; sss < in.Grid()->oSites(); sss++) , { thread_for( sss, in.Grid()->oSites(),{
Kernels::DhopDirKernel(Stencil, Umu_v, UUUmu_v, Stencil.CommBuf(), sss, sss, in_v, out_v, dir, disp); Kernels::DhopDirKernel(Stencil, Umu_v, UUUmu_v, Stencil.CommBuf(), sss, sss, in_v, out_v, dir, disp);
}); });
}; };
@@ -478,13 +478,13 @@ void ImprovedStaggeredFermion<Impl>::DhopInternalOverlappedComms(StencilImpl &st
auto out_v = out.View(); auto out_v = out.View();
if (dag == DaggerYes) { if (dag == DaggerYes) {
int sz=st.surface_list.size(); int sz=st.surface_list.size();
thread_loop( (int ss = 0; ss < sz; ss++) ,{ thread_for(ss,sz,{
int sU = st.surface_list[ss]; int sU = st.surface_list[ss];
Kernels::DhopSiteDag(st,lo,U_v,UUU_v,st.CommBuf(),1,sU,in_v,out_v,0,1); Kernels::DhopSiteDag(st,lo,U_v,UUU_v,st.CommBuf(),1,sU,in_v,out_v,0,1);
}); });
} else { } else {
int sz=st.surface_list.size(); int sz=st.surface_list.size();
thread_loop( (int ss = 0; ss < sz; ss++) ,{ thread_for(ss,sz,{
int sU = st.surface_list[ss]; int sU = st.surface_list[ss];
Kernels::DhopSite(st,lo,U_v,UUU_v,st.CommBuf(),1,sU,in_v,out_v,0,1); Kernels::DhopSite(st,lo,U_v,UUU_v,st.CommBuf(),1,sU,in_v,out_v,0,1);
}); });
@@ -519,11 +519,11 @@ void ImprovedStaggeredFermion<Impl>::DhopInternalSerialComms(StencilImpl &st, Le
auto out_v = out.View(); auto out_v = out.View();
DhopComputeTime -= usecond(); DhopComputeTime -= usecond();
if (dag == DaggerYes) { if (dag == DaggerYes) {
thread_loop( (int sss = 0; sss < in.Grid()->oSites(); sss++), { thread_for(sss, in.Grid()->oSites(),{
Kernels::DhopSiteDag(st, lo, U_v, UUU_v, st.CommBuf(), 1, sss, in_v, out_v); Kernels::DhopSiteDag(st, lo, U_v, UUU_v, st.CommBuf(), 1, sss, in_v, out_v);
}); });
} else { } else {
thread_loop( (int sss = 0; sss < in.Grid()->oSites(); sss++), { thread_for(sss, in.Grid()->oSites(),{
Kernels::DhopSite(st, lo, U_v, UUU_v, st.CommBuf(), 1, sss, in_v, out_v); Kernels::DhopSite(st, lo, U_v, UUU_v, st.CommBuf(), 1, sss, in_v, out_v);
}); });
} }