diff --git a/Grid/qcd/action/fermion/implementation/CayleyFermion5Dcache.h b/Grid/qcd/action/fermion/implementation/CayleyFermion5Dcache.h
index ca88afd8..8af3e7c0 100644
--- a/Grid/qcd/action/fermion/implementation/CayleyFermion5Dcache.h
+++ b/Grid/qcd/action/fermion/implementation/CayleyFermion5Dcache.h
@@ -127,6 +127,12 @@ CayleyFermion5D<Impl>::MooeeInv    (const FermionField &psi_i, FermionField &chi
 
   int Ls=this->Ls;
 
+  auto plee  = & lee [0];
+  auto pdee  = & dee [0];
+  auto puee  = & uee [0];
+  auto pleem = & leem[0];
+  auto pueem = & ueem[0];
+
   MooeeInvCalls++;
   MooeeInvTime-=usecond();
   uint64_t nloop = grid->oSites()/Ls;
@@ -140,27 +146,27 @@ CayleyFermion5D<Impl>::MooeeInv    (const FermionField &psi_i, FermionField &chi
     coalescedWrite(chi[ss],psi(ss)); // chi[0]=psi[0]
     for(int s=1;s<Ls;s++){
       spProj5p(tmp,chi(ss+s-1));  
-      coalescedWrite(chi[ss+s] , psi(ss+s)-lee[s-1]*tmp);
+      coalescedWrite(chi[ss+s] , psi(ss+s)-plee[s-1]*tmp);
     }
 
     // L_m^{-1} 
-    for (int s=0;s<Ls-1;s++){ // Chi[ee] = 1 - sum[s<Ls-1] -leem[s]P_- chi
+    for (int s=0;s<Ls-1;s++){ // Chi[ee] = 1 - sum[s<Ls-1] -pleem[s]P_- chi
       spProj5m(tmp,chi(ss+s));    
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - leem[s]*tmp);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pleem[s]*tmp);
     }
 
     // U_m^{-1} D^{-1}
     for (int s=0;s<Ls-1;s++){
       // Chi[s] + 1/d chi[s] 
       spProj5p(tmp,chi(ss+Ls-1)); 
-      coalescedWrite(chi[ss+s], (1.0/dee[s])*chi(ss+s)-(ueem[s]/dee[Ls-1])*tmp);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s)-(pueem[s]/pdee[Ls-1])*tmp);
     }	
-    coalescedWrite(chi[ss+Ls-1], (1.0/dee[Ls-1])*chi(ss+Ls-1));
+    coalescedWrite(chi[ss+Ls-1], (1.0/pdee[Ls-1])*chi(ss+Ls-1));
       
     // Apply U^{-1}
     for (int s=Ls-2;s>=0;s--){
       spProj5m(tmp,chi(ss+s+1));  
-      coalescedWrite(chi[ss+s], chi(ss+s) - uee[s]*tmp);
+      coalescedWrite(chi[ss+s], chi(ss+s) - puee[s]*tmp);
     }
   });
 
@@ -179,6 +185,12 @@ CayleyFermion5D<Impl>::MooeeInvDag (const FermionField &psi_i, FermionField &chi
   auto psi = psi_i.View();
   auto chi = chi_i.View();
 
+  auto plee  = & lee [0];
+  auto pdee  = & dee [0];
+  auto puee  = & uee [0];
+  auto pleem = & leem[0];
+  auto pueem = & ueem[0];
+
   assert(psi.Checkerboard() == psi.Checkerboard());
 
   MooeeInvCalls++;
@@ -195,25 +207,25 @@ CayleyFermion5D<Impl>::MooeeInvDag (const FermionField &psi_i, FermionField &chi
     coalescedWrite(chi[ss],psi(ss));
     for (int s=1;s<Ls;s++){
       spProj5m(tmp,chi(ss+s-1));
-      coalescedWrite(chi[ss+s], psi(ss+s)-conjugate(uee[s-1])*tmp);
+      coalescedWrite(chi[ss+s], psi(ss+s)-conjugate(puee[s-1])*tmp);
     }
     // U_m^{-\dagger} 
     for (int s=0;s<Ls-1;s++){
       spProj5p(tmp,chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - conjugate(ueem[s])*tmp);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - conjugate(pueem[s])*tmp);
     }
 
     // L_m^{-\dagger} D^{-dagger}
     for (int s=0;s<Ls-1;s++){
       spProj5m(tmp,chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], conjugate(1.0/dee[s])*chi(ss+s)-conjugate(leem[s]/dee[Ls-1])*tmp);
+      coalescedWrite(chi[ss+s], conjugate(1.0/pdee[s])*chi(ss+s)-conjugate(pleem[s]/pdee[Ls-1])*tmp);
     }	
-    coalescedWrite(chi[ss+Ls-1], conjugate(1.0/dee[Ls-1])*chi(ss+Ls-1));
+    coalescedWrite(chi[ss+Ls-1], conjugate(1.0/pdee[Ls-1])*chi(ss+Ls-1));
   
     // Apply L^{-dagger}
     for (int s=Ls-2;s>=0;s--){
       spProj5p(tmp,chi(ss+s+1));
-      coalescedWrite(chi[ss+s], chi(ss+s) - conjugate(lee[s])*tmp);
+      coalescedWrite(chi[ss+s], chi(ss+s) - conjugate(plee[s])*tmp);
     }
   });
   MooeeInvTime+=usecond();
diff --git a/Grid/qcd/action/fermion/implementation/DomainWallEOFAFermionCache.h b/Grid/qcd/action/fermion/implementation/DomainWallEOFAFermionCache.h
index 468bddbd..a3eca650 100644
--- a/Grid/qcd/action/fermion/implementation/DomainWallEOFAFermionCache.h
+++ b/Grid/qcd/action/fermion/implementation/DomainWallEOFAFermionCache.h
@@ -113,6 +113,13 @@ void DomainWallEOFAFermion<Impl>::MooeeInv(const FermionField& psi_i, FermionFie
   auto chi=chi_i.View();
   int Ls = this->Ls;
 
+  auto plee  = & this->lee[0];
+  auto pdee  = & this->dee[0];
+  auto puee  = & this->uee[0];
+
+  auto pleem = & this->leem[0];
+  auto pueem = & this->ueem[0];
+
   this->MooeeInvCalls++;
   this->MooeeInvTime -= usecond();
   uint64_t nloop=grid->oSites()/Ls;
@@ -126,27 +133,27 @@ void DomainWallEOFAFermion<Impl>::MooeeInv(const FermionField& psi_i, FermionFie
     coalescedWrite(chi[ss],psi(ss)); // chi[0]=psi[0]
     for(int s=1; s<Ls; s++){
       spProj5p(tmp1, chi(ss+s-1));
-      coalescedWrite(chi[ss+s], psi(ss+s) - this->lee[s-1]*tmp1);
+      coalescedWrite(chi[ss+s], psi(ss+s) - plee[s-1]*tmp1);
     }
 
     // L_m^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[ee] = 1 - sum[s<Ls-1] -leem[s]P_- chi
       spProj5m(tmp1, chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - this->leem[s]*tmp1);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pleem[s]*tmp1);
     }
 
     // U_m^{-1} D^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[s] + 1/d chi[s]
       spProj5p(tmp1, chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], (1.0/this->dee[s])*chi(ss+s) - (this->ueem[s]/this->dee[Ls])*tmp1);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s) - (pueem[s]/pdee[Ls])*tmp1);
     }
     spProj5m(tmp2, chi(ss+Ls-1));
-    coalescedWrite(chi[ss+Ls-1],(1.0/this->dee[Ls])*tmp1 + (1.0/this->dee[Ls-1])*tmp2);
+    coalescedWrite(chi[ss+Ls-1],(1.0/pdee[Ls])*tmp1 + (1.0/pdee[Ls-1])*tmp2);
 
     // Apply U^{-1}
     for(int s=Ls-2; s>=0; s--){
       spProj5m(tmp1, chi(ss+s+1));
-      coalescedWrite(chi[ss+s], chi(ss+s) - this->uee[s]*tmp1);
+      coalescedWrite(chi[ss+s], chi(ss+s) - puee[s]*tmp1);
     }
   });
   this->MooeeInvTime += usecond();
diff --git a/Grid/qcd/action/fermion/implementation/MobiusEOFAFermionCache.h b/Grid/qcd/action/fermion/implementation/MobiusEOFAFermionCache.h
index 5825c06a..650435fc 100644
--- a/Grid/qcd/action/fermion/implementation/MobiusEOFAFermionCache.h
+++ b/Grid/qcd/action/fermion/implementation/MobiusEOFAFermionCache.h
@@ -83,7 +83,8 @@ void MobiusEOFAFermion<Impl>::M5D_shift(const FermionField &psi_i, const Fermion
   auto phi = phi_i.View();
   auto chi = chi_i.View();
 
-  int shift_s = (this->pm == 1) ? (Ls-1) : 0; // s-component modified by shift operator
+  auto pm  = this->pm;
+  int shift_s = (pm == 1) ? (Ls-1) : 0; // s-component modified by shift operator
 
   assert(phi.Checkerboard() == psi.Checkerboard());
 
@@ -104,8 +105,8 @@ void MobiusEOFAFermion<Impl>::M5D_shift(const FermionField &psi_i, const Fermion
       spProj5m(tmp1, psi(idx_u));
       spProj5p(tmp2, psi(idx_l));
 
-      if(this->pm == 1){ spProj5p(tmp, psi(ss+shift_s)); }
-      else             { spProj5m(tmp, psi(ss+shift_s)); }
+      if(pm == 1){ spProj5p(tmp, psi(ss+shift_s)); }
+      else       { spProj5m(tmp, psi(ss+shift_s)); }
 
       coalescedWrite(chi[ss+s], diag[s]*phi(ss+s) + upper[s]*tmp1 +lower[s]*tmp2 + shift_coeffs[s]*tmp);
     }
@@ -169,6 +170,8 @@ void MobiusEOFAFermion<Impl>::M5Ddag_shift(const FermionField &psi_i, const Ferm
   this->M5Dcalls++;
   this->M5Dtime -= usecond();
 
+  auto pm = this->pm;
+
   int nloop = grid->oSites()/Ls;
   accelerator_for(sss,nloop,Simd::Nsimd(),{
     uint64_t ss = sss*Ls;
@@ -188,8 +191,8 @@ void MobiusEOFAFermion<Impl>::M5Ddag_shift(const FermionField &psi_i, const Ferm
 
       if(s==(Ls-1)) coalescedWrite(chi[ss+s], chi(ss+s)+ diag[s]*phi(ss+s) + upper[s]*tmp1 + lower[s]*tmp2);
       else          coalescedWrite(chi[ss+s], diag[s]*phi(ss+s) + upper[s]*tmp1 + lower[s]*tmp2);
-      if(this->pm == 1){ spProj5p(tmp, psi(ss+s)); }
-      else             { spProj5m(tmp, psi(ss+s)); }
+      if(pm == 1){ spProj5p(tmp, psi(ss+s)); }
+      else       { spProj5m(tmp, psi(ss+s)); }
 
       coalescedWrite(chi[ss+shift_s],chi(ss+shift_s)+shift_coeffs[s]*tmp);
     }
@@ -207,6 +210,12 @@ void MobiusEOFAFermion<Impl>::MooeeInv(const FermionField &psi_i, FermionField &
   auto psi = psi_i.View();
   auto chi = chi_i.View();
 
+  auto plee = & this->lee [0];
+  auto pdee = & this->dee [0];
+  auto puee = & this->uee [0];
+  auto pleem= & this->leem[0];
+  auto pueem= & this->ueem[0];
+
   if(this->shift != 0.0){ MooeeInv_shift(psi_i,chi_i); return; }
 
   this->MooeeInvCalls++;
@@ -224,26 +233,26 @@ void MobiusEOFAFermion<Impl>::MooeeInv(const FermionField &psi_i, FermionField &
     coalescedWrite(chi[ss], psi(ss)); // chi[0]=psi[0]
     for(int s=1; s<Ls; s++){
       spProj5p(tmp, chi(ss+s-1));
-      coalescedWrite(chi[ss+s], psi(ss+s) - this->lee[s-1]*tmp);
+      coalescedWrite(chi[ss+s], psi(ss+s) - plee[s-1]*tmp);
     }
 
     // L_m^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[ee] = 1 - sum[s<Ls-1] -leem[s]P_- chi
       spProj5m(tmp, chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - this->leem[s]*tmp);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pleem[s]*tmp);
     }
 
     // U_m^{-1} D^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[s] + 1/d chi[s]
       spProj5p(tmp, chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], (1.0/this->dee[s])*chi(ss+s) - (this->ueem[s]/this->dee[Ls-1])*tmp);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s) - (pueem[s]/pdee[Ls-1])*tmp);
     }
-    coalescedWrite(chi[ss+Ls-1], (1.0/this->dee[Ls-1])*chi(ss+Ls-1));
+    coalescedWrite(chi[ss+Ls-1], (1.0/pdee[Ls-1])*chi(ss+Ls-1));
 
     // Apply U^{-1}
     for(int s=Ls-2; s>=0; s--){
       spProj5m(tmp, chi(ss+s+1));
-      coalescedWrite(chi[ss+s], chi(ss+s) - this->uee[s]*tmp);
+      coalescedWrite(chi[ss+s], chi(ss+s) - puee[s]*tmp);
     }
   });
    
@@ -259,6 +268,14 @@ void MobiusEOFAFermion<Impl>::MooeeInv_shift(const FermionField &psi_i, FermionF
   auto psi = psi_i.View();
   auto chi = chi_i.View();
 
+  auto pm = this->pm;
+  auto plee = & this->lee [0];
+  auto pdee = & this->dee [0];
+  auto puee = & this->uee [0];
+  auto pleem= & this->leem[0];
+  auto pueem= & this->ueem[0];
+  auto pMooeeInv_shift_lc   = &MooeeInv_shift_lc[0];
+  auto pMooeeInv_shift_norm = &MooeeInv_shift_norm[0];
   this->MooeeInvCalls++;
   this->MooeeInvTime -= usecond();
 
@@ -272,36 +289,36 @@ void MobiusEOFAFermion<Impl>::MooeeInv_shift(const FermionField &psi_i, FermionF
 
     // Apply (L^{\prime})^{-1} and accumulate MooeeInv_shift_lc[j]*psi[j] in tmp2
     coalescedWrite(chi[ss], psi(ss)); // chi[0]=psi[0]
-    tmp2 = MooeeInv_shift_lc[0]*psi(ss);
+    tmp2 = pMooeeInv_shift_lc[0]*psi(ss);
     for(int s=1; s<Ls; s++){
       spProj5p(tmp1, chi(ss+s-1));
-      coalescedWrite(chi[ss+s], psi(ss+s) - this->lee[s-1]*tmp1);
-      tmp2 = tmp2 + MooeeInv_shift_lc[s]*psi(ss+s);
+      coalescedWrite(chi[ss+s], psi(ss+s) - plee[s-1]*tmp1);
+      tmp2 = tmp2 + pMooeeInv_shift_lc[s]*psi(ss+s);
     }
-    if(this->pm == 1){ spProj5p(tmp2_spProj, tmp2);}
-    else             { spProj5m(tmp2_spProj, tmp2); }
+    if(pm == 1){ spProj5p(tmp2_spProj, tmp2);}
+    else       { spProj5m(tmp2_spProj, tmp2); }
 
     // L_m^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[ee] = 1 - sum[s<Ls-1] -leem[s]P_- chi
       spProj5m(tmp1, chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - this->leem[s]*tmp1);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pleem[s]*tmp1);
     }
 
     // U_m^{-1} D^{-1}
     for(int s=0; s<Ls-1; s++){ // Chi[s] + 1/d chi[s]
       spProj5p(tmp1, chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], (1.0/this->dee[s])*chi(ss+s) - (this->ueem[s]/this->dee[Ls-1])*tmp1);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s) - (pueem[s]/pdee[Ls-1])*tmp1);
     }
-    // chi[ss+Ls-1] = (1.0/this->dee[Ls-1])*chi[ss+Ls-1] + MooeeInv_shift_norm[Ls-1]*tmp2_spProj;
-    coalescedWrite(chi[ss+Ls-1], (1.0/this->dee[Ls-1])*chi(ss+Ls-1));
+    // chi[ss+Ls-1] = (1.0/pdee[Ls-1])*chi[ss+Ls-1] + MooeeInv_shift_norm[Ls-1]*tmp2_spProj;
+    coalescedWrite(chi[ss+Ls-1], (1.0/pdee[Ls-1])*chi(ss+Ls-1));
     spProj5m(tmp1, chi(ss+Ls-1));
-    coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) + MooeeInv_shift_norm[Ls-1]*tmp2_spProj);
+    coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) + pMooeeInv_shift_norm[Ls-1]*tmp2_spProj);
 
     // Apply U^{-1} and add shift term
     for(int s=Ls-2; s>=0; s--){
-      coalescedWrite(chi[ss+s] , chi(ss+s) - this->uee[s]*tmp1);
+      coalescedWrite(chi[ss+s] , chi(ss+s) - puee[s]*tmp1);
       spProj5m(tmp1, chi(ss+s));
-      coalescedWrite(chi[ss+s], chi(ss+s) + MooeeInv_shift_norm[s]*tmp2_spProj);
+      coalescedWrite(chi[ss+s], chi(ss+s) + pMooeeInv_shift_norm[s]*tmp2_spProj);
     }
   });
 
@@ -319,6 +336,12 @@ void MobiusEOFAFermion<Impl>::MooeeInvDag(const FermionField &psi_i, FermionFiel
   auto psi = psi_i.View();
   auto chi = chi_i.View();
 
+  auto plee = & this->lee [0];
+  auto pdee = & this->dee [0];
+  auto puee = & this->uee [0];
+  auto pleem= & this->leem[0];
+  auto pueem= & this->ueem[0];
+
   this->MooeeInvCalls++;
   this->MooeeInvTime -= usecond();
 
@@ -334,26 +357,26 @@ void MobiusEOFAFermion<Impl>::MooeeInvDag(const FermionField &psi_i, FermionFiel
     coalescedWrite(chi[ss], psi(ss));
     for(int s=1; s<Ls; s++){
       spProj5m(tmp, chi(ss+s-1));
-      coalescedWrite(chi[ss+s], psi(ss+s) - this->uee[s-1]*tmp);
+      coalescedWrite(chi[ss+s], psi(ss+s) - puee[s-1]*tmp);
     }
     
     // U_m^{-\dag}
     for(int s=0; s<Ls-1; s++){
       spProj5p(tmp, chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - this->ueem[s]*tmp);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pueem[s]*tmp);
     }
 
     // L_m^{-\dag} D^{-dag}
     for(int s=0; s<Ls-1; s++){
       spProj5m(tmp, chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], (1.0/this->dee[s])*chi(ss+s) - (this->leem[s]/this->dee[Ls-1])*tmp);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s) - (pleem[s]/pdee[Ls-1])*tmp);
     }
-    coalescedWrite(chi[ss+Ls-1], (1.0/this->dee[Ls-1])*chi(ss+Ls-1));
+    coalescedWrite(chi[ss+Ls-1], (1.0/pdee[Ls-1])*chi(ss+Ls-1));
 
     // Apply L^{-dag}
     for(int s=Ls-2; s>=0; s--){
       spProj5p(tmp, chi(ss+s+1));
-      coalescedWrite(chi[ss+s], chi(ss+s) - this->lee[s]*tmp);
+      coalescedWrite(chi[ss+s], chi(ss+s) - plee[s]*tmp);
     }
   });
 
@@ -369,6 +392,14 @@ void MobiusEOFAFermion<Impl>::MooeeInvDag_shift(const FermionField &psi_i, Fermi
   auto chi = chi_i.View();
   int Ls = this->Ls;
 
+  auto pm = this->pm;
+  auto plee = & this->lee [0];
+  auto pdee = & this->dee [0];
+  auto puee = & this->uee [0];
+  auto pleem= & this->leem[0];
+  auto pueem= & this->ueem[0];
+  auto pMooeeInvDag_shift_lc   = &MooeeInvDag_shift_lc[0];
+  auto pMooeeInvDag_shift_norm = &MooeeInvDag_shift_norm[0];
 
   this->MooeeInvCalls++;
   this->MooeeInvTime -= usecond();
@@ -383,36 +414,36 @@ void MobiusEOFAFermion<Impl>::MooeeInvDag_shift(const FermionField &psi_i, Fermi
 
     // Apply (U^{\prime})^{-dag} and accumulate MooeeInvDag_shift_lc[j]*psi[j] in tmp2
     coalescedWrite(chi[ss], psi(ss));
-    tmp2 = MooeeInvDag_shift_lc[0]*psi(ss);
+    tmp2 = pMooeeInvDag_shift_lc[0]*psi(ss);
     for(int s=1; s<Ls; s++){
       spProj5m(tmp1, chi(ss+s-1));
-      coalescedWrite(chi[ss+s],psi(ss+s) - this->uee[s-1]*tmp1);
-      tmp2 = tmp2 + MooeeInvDag_shift_lc[s]*psi(ss+s);
+      coalescedWrite(chi[ss+s],psi(ss+s) - puee[s-1]*tmp1);
+      tmp2 = tmp2 + pMooeeInvDag_shift_lc[s]*psi(ss+s);
     }
 
-    if(this->pm == 1){ spProj5p(tmp2_spProj, tmp2);}
-    else             { spProj5m(tmp2_spProj, tmp2);}
+    if(pm == 1){ spProj5p(tmp2_spProj, tmp2);}
+    else       { spProj5m(tmp2_spProj, tmp2);}
 
     // U_m^{-\dag}
     for(int s=0; s<Ls-1; s++){
       spProj5p(tmp1, chi(ss+s));
-      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - this->ueem[s]*tmp1);
+      coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) - pueem[s]*tmp1);
     }
 
     // L_m^{-\dag} D^{-dag}
     for(int s=0; s<Ls-1; s++){
       spProj5m(tmp1, chi(ss+Ls-1));
-      coalescedWrite(chi[ss+s], (1.0/this->dee[s])*chi(ss+s) - (this->leem[s]/this->dee[Ls-1])*tmp1);
+      coalescedWrite(chi[ss+s], (1.0/pdee[s])*chi(ss+s) - (pleem[s]/pdee[Ls-1])*tmp1);
     }
-    coalescedWrite(chi[ss+Ls-1], (1.0/this->dee[Ls-1])*chi(ss+Ls-1));
+    coalescedWrite(chi[ss+Ls-1], (1.0/pdee[Ls-1])*chi(ss+Ls-1));
     spProj5p(tmp1, chi(ss+Ls-1));
-    coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) + MooeeInvDag_shift_norm[Ls-1]*tmp2_spProj);
+    coalescedWrite(chi[ss+Ls-1], chi(ss+Ls-1) + pMooeeInvDag_shift_norm[Ls-1]*tmp2_spProj);
 
     // Apply L^{-dag}
     for(int s=Ls-2; s>=0; s--){
-      coalescedWrite(chi[ss+s], chi(ss+s) - this->lee[s]*tmp1);
+      coalescedWrite(chi[ss+s], chi(ss+s) - plee[s]*tmp1);
       spProj5p(tmp1, chi(ss+s));
-      coalescedWrite(chi[ss+s], chi(ss+s) + MooeeInvDag_shift_norm[s]*tmp2_spProj);
+      coalescedWrite(chi[ss+s], chi(ss+s) + pMooeeInvDag_shift_norm[s]*tmp2_spProj);
     }
   });
 
diff --git a/Grid/qcd/action/fermion/implementation/WilsonKernelsImplementation.h b/Grid/qcd/action/fermion/implementation/WilsonKernelsImplementation.h
index 605a3aec..5eea31f5 100644
--- a/Grid/qcd/action/fermion/implementation/WilsonKernelsImplementation.h
+++ b/Grid/qcd/action/fermion/implementation/WilsonKernelsImplementation.h
@@ -382,20 +382,20 @@ void WilsonKernels<Impl>::DhopKernel(int Opt,StencilImpl &st,  DoubledGaugeField
 
    if( interior && exterior ) { 
      if (Opt == WilsonKernelsStatic::OptGeneric    ) { KERNEL_CALL(GenericDhopSiteDag); return;}
-     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDag);    return;}
 #ifndef GRID_NVCC
+     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDag);    return;}
      if (Opt == WilsonKernelsStatic::OptInlineAsm  ) {  ASM_CALL(AsmDhopSiteDag);     return;}
 #endif
    } else if( interior ) {
      if (Opt == WilsonKernelsStatic::OptGeneric    ) { KERNEL_CALL(GenericDhopSiteDagInt); return;}
-     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDagInt);    return;}
 #ifndef GRID_NVCC
+     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDagInt);    return;}
      if (Opt == WilsonKernelsStatic::OptInlineAsm  ) {  ASM_CALL(AsmDhopSiteDagInt);     return;}
 #endif
    } else if( exterior ) { 
      if (Opt == WilsonKernelsStatic::OptGeneric    ) { KERNEL_CALL(GenericDhopSiteDagExt); return;}
-     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDagExt);    return;}
 #ifndef GRID_NVCC
+     if (Opt == WilsonKernelsStatic::OptHandUnroll ) { KERNEL_CALL(HandDhopSiteDagExt);    return;}
      if (Opt == WilsonKernelsStatic::OptInlineAsm  ) {  ASM_CALL(AsmDhopSiteDagExt);     return;}
 #endif
    }