Zero changes, acceleartor on kernels and some thread loop changes

lib/qcd/utils/CovariantLaplacian.h

@@ -113,7 +113,7 @@ public:
     GaugeLinkField sum(in.Grid());
 
     for (int nu = 0; nu < Nd; nu++) {
-      sum = zero;
+      sum = Zero();
       GaugeLinkField in_nu = PeekIndex<LorentzIndex>(in, nu);
       GaugeLinkField out_nu(out.Grid());
       for (int mu = 0; mu < Nd; mu++) {
@@ -132,7 +132,7 @@ public:
     
     for (int mu = 0; mu < Nd; mu++){
       GaugeLinkField der_mu(der.Grid());
-      der_mu = zero;
+      der_mu = Zero();
       for (int nu = 0; nu < Nd; nu++){
         GaugeLinkField in_nu = PeekIndex<LorentzIndex>(in, nu);
         der_mu += U[mu] * Cshift(in_nu, mu, 1) * adj(U[mu]) * in_nu;
@@ -151,7 +151,7 @@ public:
 
     for (int mu = 0; mu < Nd; mu++) {
       GaugeLinkField der_mu(der.Grid());
-      der_mu = zero;
+      der_mu = Zero();
       for (int nu = 0; nu < Nd; nu++) {
         GaugeLinkField left_nu = PeekIndex<LorentzIndex>(left, nu);
         GaugeLinkField right_nu = PeekIndex<LorentzIndex>(right, nu);

lib/qcd/utils/GaugeFix.h

@@ -46,7 +46,7 @@ public:
     }
   }
   static void DmuAmu(const std::vector<GaugeMat> &A,GaugeMat &dmuAmu) {
-    dmuAmu=zero;
+    dmuAmu=Zero();
     for(int mu=0;mu<Nd;mu++){
       dmuAmu = dmuAmu + A[mu] - Cshift(A[mu],mu,-1);
     }
@@ -123,7 +123,7 @@ public:
     FFT theFFT((GridCartesian *)grid);
 
     LatticeComplex  Fp(grid);
-    LatticeComplex  psq(grid); psq=zero;
+    LatticeComplex  psq(grid); psq=Zero();
     LatticeComplex  pmu(grid); 
     LatticeComplex   one(grid); one = Complex(1.0,0.0);
 

lib/qcd/utils/Metric.h

@@ -63,10 +63,10 @@ public:
     // do nothing
   }
   virtual void MDeriv(const Field& in, Field& out){
-    out = zero;
+    out = Zero();
   }
   virtual void MDeriv(const Field& left, const Field& right, Field& out){
-    out = zero;
+    out = Zero();
   }
 
 };
@@ -119,10 +119,10 @@ public:
   // Correct
   RealD MomentaAction(){
     MomentaField inv(Mom.Grid());
-    inv = zero;
+    inv = Zero();
     M.Minv(Mom, inv);
     LatticeComplex Hloc(Mom.Grid());
-    Hloc = zero;
+    Hloc = Zero();
     for (int mu = 0; mu < Nd; mu++) {
       // This is not very general
       // hide in the metric
@@ -157,7 +157,7 @@ public:
     // with respect to the gauge field
     MomentaField MDer(in.Grid());
     MomentaField X(in.Grid());
-    X = zero;
+    X = Zero();
     M.Minv(in, X);  // X = G in
     M.MDeriv(X, MDer);  // MDer = U * dS/dU
     der = Implementation::projectForce(MDer);  // Ta if gauge fields
@@ -165,12 +165,12 @@ public:
   }
 
   void AuxiliaryFieldsDerivative(MomentaField& der){
-    der = zero;
+    der = Zero();
     if (1){
       // Auxiliary fields
       MomentaField der_temp(der.Grid());
       MomentaField X(der.Grid());
-      X=zero;
+      X=Zero();
       //M.M(AuxMom, X); // X = M Aux
       // Two derivative terms
       // the Mderiv need separation of left and right terms
@@ -185,7 +185,7 @@ public:
   }
 
   void DerivativeP(MomentaField& der){
-    der = zero;
+    der = Zero();
     M.Minv(Mom, der);
     // is the projection necessary here?
     // no for fields in the algebra

lib/qcd/utils/SUn.h

@@ -162,7 +162,7 @@ public:
   
   template <class cplx>
   static void generatorSigmaY(int su2Index, iSUnMatrix<cplx> &ta) {
-    ta = zero;
+    ta = Zero();
     int i1, i2;
     su2SubGroupIndex(i1, i2, su2Index);
     ta()()(i1, i2) = 1.0;
@@ -172,7 +172,7 @@ public:
   
   template <class cplx>
   static void generatorSigmaX(int su2Index, iSUnMatrix<cplx> &ta) {
-    ta = zero;
+    ta = Zero();
     cplx i(0.0, 1.0);
     int i1, i2;
     su2SubGroupIndex(i1, i2, su2Index);
@@ -184,7 +184,7 @@ public:
   template <class cplx>
   static void generatorDiagonal(int diagIndex, iSUnMatrix<cplx> &ta) {
     // diag ({1, 1, ..., 1}(k-times), -k, 0, 0, ...)
-    ta = zero;
+    ta = Zero();
     int k = diagIndex + 1;                  // diagIndex starts from 0
     for (int i = 0; i <= diagIndex; i++) {  // k iterations
       ta()()(i, i) = 1.0;
@@ -299,11 +299,11 @@ public:
     LatticeComplex ones(grid);
     ones = 1.0;
     LatticeComplex zeros(grid);
-    zeros = zero;
+    zeros = Zero();
     LatticeReal rones(grid);
     rones = 1.0;
     LatticeReal rzeros(grid);
-    rzeros = zero;
+    rzeros = Zero();
     LatticeComplex udet(grid);  // determinant of real(staple)
     LatticeInteger mask_true(grid);
     mask_true = 1;
@@ -432,13 +432,13 @@ public:
     RealD numSites = sum(rtmp);
     RealD numAccepted;
     LatticeInteger Accepted(grid);
-    Accepted = zero;
+    Accepted = Zero();
     LatticeInteger newlyAccepted(grid);
 
     std::vector<LatticeReal> xr(4, grid);
     std::vector<LatticeReal> a(4, grid);
     LatticeReal d(grid);
-    d = zero;
+    d = Zero();
     LatticeReal alpha(grid);
 
     //    std::cout<<GridLogMessage<<"xi "<<xi <<std::endl;
@@ -475,7 +475,7 @@ public:
       LatticeInteger ione(grid);
       ione = 1;
       LatticeInteger izero(grid);
-      izero = zero;
+      izero = Zero();
 
       newlyAccepted = where(xrsq < thresh, ione, izero);
       Accepted = where(newlyAccepted, newlyAccepted, Accepted);
@@ -490,7 +490,7 @@ public:
     } while ((numAccepted < numSites) && (hit < nheatbath));
 
     // G. Set a0 = 1 - d;
-    a[0] = zero;
+    a[0] = Zero();
     a[0] = where(wheremask, 1.0 - d, a[0]);
 
     //////////////////////////////////////////
@@ -528,7 +528,7 @@ public:
     // Debug Checks
     // SU2 check
     LatticeSU2Matrix check(grid);  // rotated matrix after hb
-    u = zero;
+    u = Zero();
     check = ua * adj(ua) - 1.0;
     check = where(Accepted, check, u);
     assert(norm2(check) < 1.0e-4);
@@ -538,7 +538,7 @@ public:
     assert(norm2(check) < 1.0e-4);
 
     LatticeMatrix Vcheck(grid);
-    Vcheck = zero;
+    Vcheck = Zero();
     Vcheck = where(Accepted, V * adj(V) - 1.0, Vcheck);
     //    std::cout<<GridLogMessage << "SU3 check " <<norm2(Vcheck)<<std::endl;
     assert(norm2(Vcheck) < 1.0e-4);
@@ -622,7 +622,7 @@ public:
     ComplexD cone(1.0, 0.0);
     MatrixType ta;
 
-    lie = zero;
+    lie = Zero();
     for (int a = 0; a < AdjointDimension; a++) {
       random(pRNG, ca);
 
@@ -647,7 +647,7 @@ public:
     Complex ci(0.0, scale);
     Matrix ta;
 
-    out = zero;
+    out = Zero();
     for (int a = 0; a < AdjointDimension; a++) {
       gaussian(pRNG, ca);
       generator(a, ta);
@@ -665,7 +665,7 @@ public:
     LatticeMatrix la(grid);
     Matrix ta;
 
-    out = zero;
+    out = Zero();
     for (int a = 0; a < AdjointDimension; a++) {
       generator(a, ta);
       la = peekColour(h, a) * timesI(ta) * scale;
@@ -708,7 +708,7 @@ public:
   // inverse operation: FundamentalLieAlgebraMatrix
   static void projectOnAlgebra(LatticeAlgebraVector &h_out, const LatticeMatrix &in, Real scale = 1.0) {
     conformable(h_out, in);
-    h_out = zero;
+    h_out = Zero();
     Matrix Ta;
 
     for (int a = 0; a < AdjointDimension; a++) {

lib/qcd/utils/SUnAdjoint.h

@@ -61,7 +61,7 @@ public:
   static void generator(int Index, iSUnAdjointMatrix<cplx> &iAdjTa) {
     // returns i(T_Adj)^index necessary for the projectors
     // see definitions above
-    iAdjTa = zero;
+    iAdjTa = Zero();
     Vector<typename SU<ncolour>::template iSUnMatrix<cplx> > ta(ncolour * ncolour - 1);
     typename SU<ncolour>::template iSUnMatrix<cplx> tmp;
 
@@ -118,7 +118,7 @@ public:
     LatticeAdjMatrix la(grid);
     AMatrix iTa;
 
-    out = zero;
+    out = Zero();
     for (int a = 0; a < Dimension; a++) {
       generator(a, iTa);
       la = peekColour(h, a) * iTa;
@@ -130,7 +130,7 @@ public:
   // Projects the algebra components a lattice matrix (of dimension ncol*ncol -1 )
   static void projectOnAlgebra(typename SU<ncolour>::LatticeAlgebraVector &h_out, const LatticeAdjMatrix &in, Real scale = 1.0) {
     conformable(h_out, in);
-    h_out = zero;
+    h_out = Zero();
     AMatrix iTa;
     Real coefficient = - 1.0/(ncolour) * scale;// 1/Nc for the normalization of the trace in the adj rep
 
@@ -145,7 +145,7 @@ public:
   static void projector(typename SU<ncolour>::LatticeAlgebraVector &h_out, const LatticeAdjMatrix &in, Real scale = 1.0) {
     conformable(h_out, in);
     static std::vector<AMatrix> iTa(Dimension);  // to store the generators
-    h_out = zero;
+    h_out = Zero();
     static bool precalculated = false; 
     if (!precalculated){
       precalculated = true;

lib/qcd/utils/SUnTwoIndex.h

@@ -71,7 +71,7 @@ public:
   static void base(int Index, iSUnMatrix<cplx> &eij) {
     // returns (e)^(ij)_{kl} necessary for change of base U_F -> U_R
     assert(Index < NumGenerators);
-    eij = zero;
+    eij = Zero();
 
     // for the linearisation of the 2 indexes 
     static int a[ncolour * (ncolour - 1) / 2][2]; // store the a <-> i,j
@@ -97,14 +97,14 @@ public:
 
   template <class cplx>
   static void baseDiagonal(int Index, iSUnMatrix<cplx> &eij) {
-    eij = zero;
+    eij = Zero();
     eij()()(Index - ncolour * (ncolour - 1) / 2,
             Index - ncolour * (ncolour - 1) / 2) = 1.0;
   }
 
   template <class cplx>
   static void baseOffDiagonal(int i, int j, iSUnMatrix<cplx> &eij) {
-    eij = zero;
+    eij = Zero();
     for (int k = 0; k < ncolour; k++)
       for (int l = 0; l < ncolour; l++)
         eij()()(l, k) = delta(i, k) * delta(j, l) +
@@ -130,7 +130,7 @@ public:
 								ncolour * ncolour - 1);
     Vector<typename SU<ncolour>::template iSUnMatrix<cplx> > eij(Dimension);
     typename SU<ncolour>::template iSUnMatrix<cplx> tmp;
-    i2indTa = zero;
+    i2indTa = Zero();
     
     for (int a = 0; a < ncolour * ncolour - 1; a++)
       SU<ncolour>::generator(a, ta[a]);
@@ -203,7 +203,7 @@ public:
     LatticeTwoIndexMatrix la(grid);
     TIMatrix i2indTa;
 
-    out = zero;
+    out = Zero();
     for (int a = 0; a < ncolour * ncolour - 1; a++) {
       generator(a, i2indTa);
       la = peekColour(h, a) * i2indTa;
@@ -218,7 +218,7 @@ public:
 			       typename SU<ncolour>::LatticeAlgebraVector &h_out,
 			       const LatticeTwoIndexMatrix &in, Real scale = 1.0) {
     conformable(h_out, in);
-    h_out = zero;
+    h_out = Zero();
     TIMatrix i2indTa;
     Real coefficient = -2.0 / (ncolour + 2 * S) * scale;
     // 2/(Nc +/- 2) for the normalization of the trace in the two index rep
@@ -236,7 +236,7 @@ public:
     conformable(h_out, in);
     // to store the generators
     static std::vector<TIMatrix> i2indTa(ncolour * ncolour -1); 
-    h_out = zero;
+    h_out = Zero();
     static bool precalculated = false;
     if (!precalculated) {
       precalculated = true;

lib/qcd/utils/WilsonLoops.h

@@ -86,7 +86,7 @@ public:
   static void sitePlaquette(ComplexField &Plaq,
                             const std::vector<GaugeMat> &U) {
     ComplexField sitePlaq(U[0].Grid());
-    Plaq = zero;
+    Plaq = Zero();
     for (int mu = 1; mu < Nd; mu++) {
       for (int nu = 0; nu < mu; nu++) {
         traceDirPlaquette(sitePlaq, U, mu, nu);
@@ -130,7 +130,7 @@ public:
     std::vector<GaugeMat> U(Nd, Umu.Grid());
 
     ComplexField Tr(Umu.Grid());
-    Tr = zero;
+    Tr = Zero();
     for (int mu = 0; mu < Nd; mu++) {
       U[mu] = PeekIndex<LorentzIndex>(Umu, mu);
       Tr = Tr + trace(U[mu]);
@@ -156,7 +156,7 @@ public:
     for (int d = 0; d < Nd; d++) {
       U[d] = PeekIndex<LorentzIndex>(Umu, d);
     }
-    staple = zero;
+    staple = Zero();
 
     if (nu != mu) {
 
@@ -197,7 +197,7 @@ public:
       // this operation is taking too much time
       U[d] = PeekIndex<LorentzIndex>(Umu, d);
     }
-    staple = zero;
+    staple = Zero();
     GaugeMat tmp1(grid);
     GaugeMat tmp2(grid);
 
@@ -225,7 +225,7 @@ public:
     for (int d = 0; d < Nd; d++) {
       U[d] = PeekIndex<LorentzIndex>(Umu, d);
     }
-    staple = zero;
+    staple = Zero();
 
     for (int nu = 0; nu < Nd; nu++) {
 
@@ -385,7 +385,7 @@ public:
   static void siteRectangle(ComplexField &Rect,
                             const std::vector<GaugeMat> &U) {
     ComplexField siteRect(U[0].Grid());
-    Rect = zero;
+    Rect = Zero();
     for (int mu = 1; mu < Nd; mu++) {
       for (int nu = 0; nu < mu; nu++) {
         traceDirRectangle(siteRect, U, mu, nu);
@@ -443,7 +443,7 @@ public:
   static void RectStapleOptimised(GaugeMat &Stap, std::vector<GaugeMat> &U2,
                                   std::vector<GaugeMat> &U, int mu) {
 
-    Stap = zero;
+    Stap = Zero();
 
     GridBase *grid = U[0].Grid();
 
@@ -529,7 +529,7 @@ public:
       U[d] = PeekIndex<LorentzIndex>(Umu, d);
     }
 
-    Stap = zero;
+    Stap = Zero();
 
     for (int nu = 0; nu < Nd; nu++) {
       if (nu != mu) {