portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-06-18 07:47:06 +01:00
Code Releases Activity

View introduction to prepare for accelerator offload.

Browse Source 
Probably same problem exists for stencil object
This commit is contained in:
paboyle
2018-03-04 16:38:08 +00:00
parent 442b0b406c
commit 3277bda130
19 changed files with 595 additions and 457 deletions
Download Patch File Download Diff File Expand all files Collapse all files

72
lib/qcd/action/fermion/WilsonFermion.cc
View File

@ -216,8 +216,11 @@ void WilsonFermion<Impl>::DerivInternal(StencilImpl &st, DoubledGaugeField &U,
////////////////////////
// Call the single hop
////////////////////////
auto U_v = U.View();
auto B_v = B.View();
auto Btilde_v = Btilde.View();
thread_loop( (int sss = 0; sss < B.Grid()->oSites(); sss++) ,{
Kernels::DhopDirK(st, U, st.CommBuf(), sss, sss, B, Btilde, mu, gamma);
Kernels::DhopDirK(st, U_v, st.CommBuf(), sss, sss, B_v, Btilde_v, mu, gamma);
});
//////////////////////////////////////////////////
@ -316,9 +319,11 @@ void WilsonFermion<Impl>::DhopDirDisp(const FermionField &in, FermionField &out,
Compressor compressor(dag);
Stencil.HaloExchange(in, compressor);
auto in_v = in.View();
auto out_v = in.View();
auto Umu_v = Umu.View();
thread_loop( (int sss = 0; sss < in.Grid()->oSites(); sss++) ,{
Kernels::DhopDirK(Stencil, Umu, Stencil.CommBuf(), sss, sss, in, out, dirdisp, gamma);
Kernels::DhopDirK(Stencil, Umu_v, Stencil.CommBuf(), sss, sss, in_v, out_v, dirdisp, gamma);
});
};
@ -333,13 +338,16 @@ void WilsonFermion<Impl>::DhopInternal(StencilImpl &st, LebesgueOrder &lo,
st.HaloExchange(in, compressor);
int Opt = WilsonKernelsStatic::Opt;
auto U_v = U.View();
auto in_v = in.View();
auto out_v= out.View();
if (dag == DaggerYes) {
accelerator_loop( sss,in, {
Kernels::DhopSiteDag(Opt,st, lo, U, st.CommBuf(), sss, sss, 1, 1, in, out);
accelerator_loop( sss,in_v, {
Kernels::DhopSiteDag(Opt,st, lo, U_v, st.CommBuf(), sss, sss, 1, 1, in_v, out_v);
});
} else {
accelerator_loop( sss,in, {
Kernels::DhopSite(Opt,st, lo, U, st.CommBuf(), sss, sss, 1, 1, in, out);
accelerator_loop( sss,in_v, {
Kernels::DhopSite(Opt,st, lo, U_v, st.CommBuf(), sss, sss, 1, 1, in_v, out_v);
});
}
};
@ -367,15 +375,21 @@ void WilsonFermion<Impl>::ContractConservedCurrent(PropagatorField &q_in_1,
// Inefficient comms method but not performance critical.
tmp1 = Cshift(q_in_1, mu, 1);
tmp2 = Cshift(q_in_2, mu, 1);
auto tmp1_v = tmp1.View();
auto tmp2_v = tmp2.View();
auto q_in_1_v=q_in_1.View();
auto q_in_2_v=q_in_2.View();
auto q_out_v = q_out.View();
auto Umu_v = Umu.View();
thread_loop( (unsigned int sU = 0; sU < Umu.Grid()->oSites(); ++sU), {
Kernels::ContractConservedCurrentSiteFwd(tmp1[sU],
q_in_2[sU],
q_out[sU],
Umu, sU, mu);
Kernels::ContractConservedCurrentSiteBwd(q_in_1[sU],
tmp2[sU],
q_out[sU],
Umu, sU, mu);
Kernels::ContractConservedCurrentSiteFwd(tmp1_v[sU],
q_in_2_v[sU],
q_out_v[sU],
Umu_v, sU, mu);
Kernels::ContractConservedCurrentSiteBwd(q_in_1_v[sU],
tmp2_v[sU],
q_out_v[sU],
Umu_v, sU, mu);
});
}
@ -415,34 +429,40 @@ void WilsonFermion<Impl>::SeqConservedCurrent(PropagatorField &q_in,
tmp = ph*q_in;
tmpBwd = Cshift(tmp, mu, -1);
auto coords_v = coords.View();
auto tmpFwd_v = tmpFwd.View();
auto tmpBwd_v = tmpBwd.View();
auto Umu_v = Umu.View();
auto q_out_v = q_out.View();
thread_loop( (unsigned int sU = 0; sU < Umu.Grid()->oSites(); ++sU), {
// Compute the sequential conserved current insertion only if our simd
// object contains a timeslice we need.
vInteger t_mask = ((coords[sU] >= tmin) &&
(coords[sU] <= tmax));
vInteger t_mask = ((coords_v[sU] >= tmin) &&
(coords_v[sU] <= tmax));
Integer timeSlices = Reduce(t_mask);
if (timeSlices > 0) {
Kernels::SeqConservedCurrentSiteFwd(tmpFwd[sU],
q_out[sU],
Umu, sU, mu, t_mask);
Kernels::SeqConservedCurrentSiteFwd(tmpFwd_v[sU],
q_out_v[sU],
Umu_v, sU, mu, t_mask);
}
// Repeat for backward direction.
t_mask = ((coords[sU] >= (tmin + tshift)) &&
(coords[sU] <= (tmax + tshift)));
t_mask = ((coords_v[sU] >= (tmin + tshift)) &&
(coords_v[sU] <= (tmax + tshift)));
//if tmax = LLt-1 (last timeslice) include timeslice 0 if the time is shifted (mu=3)
unsigned int t0 = 0;
if((tmax==LLt-1) && (tshift==1)) t_mask = (t_mask || (coords[sU] == t0 ));
if((tmax==LLt-1) && (tshift==1)) t_mask = (t_mask || (coords_v[sU] == t0 ));
timeSlices = Reduce(t_mask);
if (timeSlices > 0) {
Kernels::SeqConservedCurrentSiteBwd(tmpBwd[sU],
q_out[sU],
Umu, sU, mu, t_mask);
Kernels::SeqConservedCurrentSiteBwd(tmpBwd_v[sU],
q_out_v[sU],
Umu_v, sU, mu, t_mask);
}
});
}