From 3ce10aa97592c412e22ce51c1ace479f9a0f4357 Mon Sep 17 00:00:00 2001
From: paboyle <paboyle@ph.ed.ac.uk>
Date: Thu, 10 Dec 2015 22:55:00 +0000
Subject: [PATCH] Fix a regression failure on Mobius; chroma regression added

---
 benchmarks/Benchmark_dwf.cc                 |  8 ++---
 benchmarks/Benchmark_wilson.cc              |  8 ++---
 configure.ac                                | 16 +++++++++
 lib/Config.h.in                             |  3 ++
 lib/Make.inc                                |  4 +--
 lib/PerfCount.cc                            |  4 +++
 lib/PerfCount.h                             |  1 +
 lib/Simd.h                                  |  3 +-
 lib/qcd/action/fermion/CayleyFermion5D.cc   |  5 ++-
 lib/qcd/action/fermion/WilsonKernels.cc     | 21 +++++++++---
 lib/qcd/action/fermion/WilsonKernels.h      | 31 +++---------------
 lib/qcd/action/fermion/WilsonKernelsHand.cc | 36 +++++++++++++++++++--
 lib/simd/Grid_avx.h                         |  5 +--
 scripts/filelist                            | 22 +++++++++++--
 tests/Make.inc                              | 10 ++++--
 tests/Makefile.am                           |  7 ++++
 tests/Test_cayley_cg.cc                     |  3 +-
 tests/Test_dwf_even_odd.cc                  |  2 +-
 18 files changed, 132 insertions(+), 57 deletions(-)

diff --git a/benchmarks/Benchmark_dwf.cc b/benchmarks/Benchmark_dwf.cc
index f86fd072..1c871e72 100644
--- a/benchmarks/Benchmark_dwf.cc
+++ b/benchmarks/Benchmark_dwf.cc
@@ -68,11 +68,11 @@ int main (int argc, char ** argv)
     for(int mu=0;mu<Nd;mu++){
 
       tmp = U[mu]*Cshift(src,mu+1,1);
-      ref=ref + tmp + Gamma(Gmu[mu])*tmp;
+      ref=ref + tmp - Gamma(Gmu[mu])*tmp;
 
       tmp =adj(U[mu])*src;
       tmp =Cshift(tmp,mu+1,-1);
-      ref=ref + tmp - Gamma(Gmu[mu])*tmp;
+      ref=ref + tmp + Gamma(Gmu[mu])*tmp;
     }
     ref = -0.5*ref;
   }
@@ -111,13 +111,13 @@ int main (int argc, char ** argv)
       //    ref =  src - Gamma(Gamma::GammaX)* src ; // 1+gamma_x
       tmp = U[mu]*Cshift(src,mu+1,1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
 
       tmp =adj(U[mu])*src;
       tmp =Cshift(tmp,mu+1,-1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
     }
     ref = -0.5*ref;
diff --git a/benchmarks/Benchmark_wilson.cc b/benchmarks/Benchmark_wilson.cc
index fc29e34b..eab0fdd2 100644
--- a/benchmarks/Benchmark_wilson.cc
+++ b/benchmarks/Benchmark_wilson.cc
@@ -75,13 +75,13 @@ int main (int argc, char ** argv)
       //    ref =  src + Gamma(Gamma::GammaX)* src ; // 1-gamma_x
       tmp = U[mu]*Cshift(src,mu,1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
 
       tmp =adj(U[mu])*src;
       tmp =Cshift(tmp,mu,-1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
     }
   }
@@ -124,13 +124,13 @@ int main (int argc, char ** argv)
       //    ref =  src - Gamma(Gamma::GammaX)* src ; // 1+gamma_x
       tmp = U[mu]*Cshift(src,mu,1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
 
       tmp =adj(U[mu])*src;
       tmp =Cshift(tmp,mu,-1);
       for(int i=0;i<ref._odata.size();i++){
-	ref._odata[i]+= tmp._odata[i] + Gamma(Gmu[mu])*tmp._odata[i]; ;
+	ref._odata[i]+= tmp._odata[i] - Gamma(Gmu[mu])*tmp._odata[i]; ;
       }
     }
   }
diff --git a/configure.ac b/configure.ac
index 1fcd4f8f..c8fd8b1a 100644
--- a/configure.ac
+++ b/configure.ac
@@ -164,6 +164,21 @@ esac
 AM_CONDITIONAL(BUILD_COMMS_MPI,[ test "X${ac_COMMS}X" == "XmpiX" ])
 AM_CONDITIONAL(BUILD_COMMS_NONE,[ test "X${ac_COMMS}X" == "XnoneX" ])
 
+AC_ARG_ENABLE([chroma],[AC_HELP_STRING([--enable-chroma],[Expect chroma compiled under c++11 ])],ac_CHROMA=yes,ac_CHROMA=no)
+
+case ${ac_CHROMA} in
+     yes)
+       echo Enabling tests regressing to Chroma
+     ;;
+     no)
+       echo Disabling tests regressing to Chroma
+     ;;
+     *)
+     AC_MSG_ERROR([${ac_CHROMA} unsupported --enable-chroma option]); 
+     ;;
+esac
+
+AM_CONDITIONAL(BUILD_CHROMA_REGRESSION,[ test "X${ac_CHROMA}X" == "XyesX" ])
 
 ###################################################################
 # Checks for doxygen support
@@ -184,6 +199,7 @@ echo :::::::::::::::::::::::::::::::::::::::::::
 AC_CONFIG_FILES(Makefile)
 AC_CONFIG_FILES(lib/Makefile)
 AC_CONFIG_FILES(tests/Makefile)
+AC_CONFIG_FILES(tests/qdpxx/Makefile)
 AC_CONFIG_FILES(benchmarks/Makefile)
 AC_OUTPUT
 
diff --git a/lib/Config.h.in b/lib/Config.h.in
index 2cc9106a..4704273d 100644
--- a/lib/Config.h.in
+++ b/lib/Config.h.in
@@ -137,6 +137,9 @@
 /* Define to the one symbol short name of this package. */
 #undef PACKAGE_TARNAME
 
+/* Define to the home page for this package. */
+#undef PACKAGE_URL
+
 /* Define to the version of this package. */
 #undef PACKAGE_VERSION
 
diff --git a/lib/Make.inc b/lib/Make.inc
index a13d8f27..de647bcb 100644
--- a/lib/Make.inc
+++ b/lib/Make.inc
@@ -1,4 +1,4 @@
 
-HFILES=./algorithms/approx/bigfloat.h ./algorithms/approx/bigfloat_double.h ./algorithms/approx/Chebyshev.h ./algorithms/approx/MultiShiftFunction.h ./algorithms/approx/Remez.h ./algorithms/approx/Zolotarev.h ./algorithms/CoarsenedMatrix.h ./algorithms/iterative/AdefGeneric.h ./algorithms/iterative/ConjugateGradient.h ./algorithms/iterative/ConjugateGradientMultiShift.h ./algorithms/iterative/ConjugateResidual.h ./algorithms/iterative/ImplicitlyRestartedLanczos.h ./algorithms/iterative/MatrixUtils.h ./algorithms/iterative/NormalEquations.h ./algorithms/iterative/PrecConjugateResidual.h ./algorithms/iterative/PrecGeneralisedConjugateResidual.h ./algorithms/iterative/SchurRedBlack.h ./algorithms/LinearOperator.h ./algorithms/Preconditioner.h ./algorithms/SparseMatrix.h ./Algorithms.h ./AlignedAllocator.h ./cartesian/Cartesian_base.h ./cartesian/Cartesian_full.h ./cartesian/Cartesian_red_black.h ./Cartesian.h ./communicator/Communicator_base.h ./Communicator.h ./Config.h ./cshift/Cshift_common.h ./cshift/Cshift_mpi.h ./cshift/Cshift_none.h ./Cshift.h ./Grid.h ./Init.h ./lattice/Lattice_arith.h ./lattice/Lattice_base.h ./lattice/Lattice_comparison.h ./lattice/Lattice_comparison_utils.h ./lattice/Lattice_conformable.h ./lattice/Lattice_coordinate.h ./lattice/Lattice_ET.h ./lattice/Lattice_local.h ./lattice/Lattice_overload.h ./lattice/Lattice_peekpoke.h ./lattice/Lattice_reality.h ./lattice/Lattice_reduction.h ./lattice/Lattice_rng.h ./lattice/Lattice_trace.h ./lattice/Lattice_transfer.h ./lattice/Lattice_transpose.h ./lattice/Lattice_unary.h ./lattice/Lattice_where.h ./Lattice.h ./Log.h ./Old/Tensor_peek.h ./Old/Tensor_poke.h ./parallelIO/BinaryIO.h ./parallelIO/NerscIO.h ./pugixml/pugixml.h ./qcd/action/ActionBase.h ./qcd/action/ActionParams.h ./qcd/action/Actions.h ./qcd/action/fermion/CayleyFermion5D.h ./qcd/action/fermion/ContinuedFractionFermion5D.h ./qcd/action/fermion/DomainWallFermion.h ./qcd/action/fermion/FermionOperator.h ./qcd/action/fermion/FermionOperatorImpl.h ./qcd/action/fermion/g5HermitianLinop.h ./qcd/action/fermion/MobiusFermion.h ./qcd/action/fermion/MobiusZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonCayleyTanhFermion.h ./qcd/action/fermion/OverlapWilsonCayleyZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonContfracTanhFermion.h ./qcd/action/fermion/OverlapWilsonContfracZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonPartialFractionTanhFermion.h ./qcd/action/fermion/OverlapWilsonPartialFractionZolotarevFermion.h ./qcd/action/fermion/PartialFractionFermion5D.h ./qcd/action/fermion/ScaledShamirFermion.h ./qcd/action/fermion/ShamirZolotarevFermion.h ./qcd/action/fermion/WilsonCompressor.h ./qcd/action/fermion/WilsonFermion.h ./qcd/action/fermion/WilsonFermion5D.h ./qcd/action/fermion/WilsonKernels.h ./qcd/action/gauge/WilsonGaugeAction.h ./qcd/action/pseudofermion/EvenOddSchurDifferentiable.h ./qcd/action/pseudofermion/OneFlavourEvenOddRational.h ./qcd/action/pseudofermion/OneFlavourEvenOddRationalRatio.h ./qcd/action/pseudofermion/OneFlavourRational.h ./qcd/action/pseudofermion/OneFlavourRationalRatio.h ./qcd/action/pseudofermion/TwoFlavour.h ./qcd/action/pseudofermion/TwoFlavourEvenOdd.h ./qcd/action/pseudofermion/TwoFlavourEvenOddRatio.h ./qcd/action/pseudofermion/TwoFlavourRatio.h ./qcd/hmc/HMC.h ./qcd/hmc/integrators/Integrator.h ./qcd/hmc/integrators/Integrator_algorithm.h ./qcd/QCD.h ./qcd/spin/Dirac.h ./qcd/spin/TwoSpinor.h ./qcd/utils/CovariantCshift.h ./qcd/utils/LinalgUtils.h ./qcd/utils/SpaceTimeGrid.h ./qcd/utils/SUn.h ./qcd/utils/WilsonLoops.h ./serialisation/BaseIO.h ./serialisation/BinaryIO.h ./serialisation/MacroMagic.h ./serialisation/Serialisation.h ./serialisation/TextIO.h ./serialisation/XmlIO.h ./simd/Grid_avx.h ./simd/Grid_avx512.h ./simd/Grid_empty.h ./simd/Grid_imci.h ./simd/Grid_neon.h ./simd/Grid_qpx.h ./simd/Grid_sse4.h ./simd/Grid_vector_types.h ./simd/Grid_vector_unops.h ./Simd.h ./stencil/Lebesgue.h ./Stencil.h ./tensors/Tensor_arith.h ./tensors/Tensor_arith_add.h ./tensors/Tensor_arith_mac.h ./tensors/Tensor_arith_mul.h ./tensors/Tensor_arith_scalar.h ./tensors/Tensor_arith_sub.h ./tensors/Tensor_class.h ./tensors/Tensor_determinant.h ./tensors/Tensor_exp.h ./tensors/Tensor_extract_merge.h ./tensors/Tensor_index.h ./tensors/Tensor_inner.h ./tensors/Tensor_logical.h ./tensors/Tensor_outer.h ./tensors/Tensor_reality.h ./tensors/Tensor_Ta.h ./tensors/Tensor_trace.h ./tensors/Tensor_traits.h ./tensors/Tensor_transpose.h ./tensors/Tensor_unary.h ./Tensors.h ./Threads.h ./Timer.h
+HFILES=./algorithms/approx/bigfloat.h ./algorithms/approx/bigfloat_double.h ./algorithms/approx/Chebyshev.h ./algorithms/approx/MultiShiftFunction.h ./algorithms/approx/Remez.h ./algorithms/approx/Zolotarev.h ./algorithms/CoarsenedMatrix.h ./algorithms/iterative/AdefGeneric.h ./algorithms/iterative/ConjugateGradient.h ./algorithms/iterative/ConjugateGradientMultiShift.h ./algorithms/iterative/ConjugateResidual.h ./algorithms/iterative/DenseMatrix.h ./algorithms/iterative/EigenSort.h ./algorithms/iterative/Francis.h ./algorithms/iterative/Householder.h ./algorithms/iterative/ImplicitlyRestartedLanczos.h ./algorithms/iterative/Matrix.h ./algorithms/iterative/MatrixUtils.h ./algorithms/iterative/NormalEquations.h ./algorithms/iterative/PrecConjugateResidual.h ./algorithms/iterative/PrecGeneralisedConjugateResidual.h ./algorithms/iterative/SchurRedBlack.h ./algorithms/LinearOperator.h ./algorithms/Preconditioner.h ./algorithms/SparseMatrix.h ./Algorithms.h ./AlignedAllocator.h ./cartesian/Cartesian_base.h ./cartesian/Cartesian_full.h ./cartesian/Cartesian_red_black.h ./Cartesian.h ./communicator/Communicator_base.h ./Communicator.h ./cshift/Cshift_common.h ./cshift/Cshift_mpi.h ./cshift/Cshift_none.h ./Cshift.h ./Grid.h ./Init.h ./lattice/Lattice_arith.h ./lattice/Lattice_base.h ./lattice/Lattice_comparison.h ./lattice/Lattice_comparison_utils.h ./lattice/Lattice_conformable.h ./lattice/Lattice_coordinate.h ./lattice/Lattice_ET.h ./lattice/Lattice_local.h ./lattice/Lattice_overload.h ./lattice/Lattice_peekpoke.h ./lattice/Lattice_reality.h ./lattice/Lattice_reduction.h ./lattice/Lattice_rng.h ./lattice/Lattice_trace.h ./lattice/Lattice_transfer.h ./lattice/Lattice_transpose.h ./lattice/Lattice_unary.h ./lattice/Lattice_where.h ./Lattice.h ./Log.h ./Old/Tensor_peek.h ./Old/Tensor_poke.h ./parallelIO/BinaryIO.h ./parallelIO/NerscIO.h ./PerfCount.h ./pugixml/pugixml.h ./qcd/action/ActionBase.h ./qcd/action/ActionParams.h ./qcd/action/Actions.h ./qcd/action/fermion/CayleyFermion5D.h ./qcd/action/fermion/ContinuedFractionFermion5D.h ./qcd/action/fermion/DomainWallFermion.h ./qcd/action/fermion/FermionOperator.h ./qcd/action/fermion/FermionOperatorImpl.h ./qcd/action/fermion/g5HermitianLinop.h ./qcd/action/fermion/MobiusFermion.h ./qcd/action/fermion/MobiusZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonCayleyTanhFermion.h ./qcd/action/fermion/OverlapWilsonCayleyZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonContfracTanhFermion.h ./qcd/action/fermion/OverlapWilsonContfracZolotarevFermion.h ./qcd/action/fermion/OverlapWilsonPartialFractionTanhFermion.h ./qcd/action/fermion/OverlapWilsonPartialFractionZolotarevFermion.h ./qcd/action/fermion/PartialFractionFermion5D.h ./qcd/action/fermion/ScaledShamirFermion.h ./qcd/action/fermion/ShamirZolotarevFermion.h ./qcd/action/fermion/WilsonCompressor.h ./qcd/action/fermion/WilsonFermion.h ./qcd/action/fermion/WilsonFermion5D.h ./qcd/action/fermion/WilsonKernels.h ./qcd/action/gauge/WilsonGaugeAction.h ./qcd/action/pseudofermion/EvenOddSchurDifferentiable.h ./qcd/action/pseudofermion/OneFlavourEvenOddRational.h ./qcd/action/pseudofermion/OneFlavourEvenOddRationalRatio.h ./qcd/action/pseudofermion/OneFlavourRational.h ./qcd/action/pseudofermion/OneFlavourRationalRatio.h ./qcd/action/pseudofermion/TwoFlavour.h ./qcd/action/pseudofermion/TwoFlavourEvenOdd.h ./qcd/action/pseudofermion/TwoFlavourEvenOddRatio.h ./qcd/action/pseudofermion/TwoFlavourRatio.h ./qcd/hmc/HMC.h ./qcd/hmc/integrators/Integrator.h ./qcd/hmc/integrators/Integrator_algorithm.h ./qcd/QCD.h ./qcd/spin/Dirac.h ./qcd/spin/TwoSpinor.h ./qcd/utils/CovariantCshift.h ./qcd/utils/LinalgUtils.h ./qcd/utils/SpaceTimeGrid.h ./qcd/utils/SUn.h ./qcd/utils/WilsonLoops.h ./serialisation/BaseIO.h ./serialisation/BinaryIO.h ./serialisation/MacroMagic.h ./serialisation/Serialisation.h ./serialisation/TextIO.h ./serialisation/XmlIO.h ./simd/Avx512Asm.h ./simd/Grid_avx.h ./simd/Grid_avx512.h ./simd/Grid_empty.h ./simd/Grid_imci.h ./simd/Grid_neon.h ./simd/Grid_qpx.h ./simd/Grid_sse4.h ./simd/Grid_vector_types.h ./simd/Grid_vector_unops.h ./Simd.h ./stencil/Lebesgue.h ./Stencil.h ./tensors/Tensor_arith.h ./tensors/Tensor_arith_add.h ./tensors/Tensor_arith_mac.h ./tensors/Tensor_arith_mul.h ./tensors/Tensor_arith_scalar.h ./tensors/Tensor_arith_sub.h ./tensors/Tensor_class.h ./tensors/Tensor_determinant.h ./tensors/Tensor_exp.h ./tensors/Tensor_extract_merge.h ./tensors/Tensor_index.h ./tensors/Tensor_inner.h ./tensors/Tensor_logical.h ./tensors/Tensor_outer.h ./tensors/Tensor_reality.h ./tensors/Tensor_Ta.h ./tensors/Tensor_trace.h ./tensors/Tensor_traits.h ./tensors/Tensor_transpose.h ./tensors/Tensor_unary.h ./Tensors.h ./Threads.h ./Timer.h
 
-CCFILES=./algorithms/approx/MultiShiftFunction.cc ./algorithms/approx/Remez.cc ./algorithms/approx/Zolotarev.cc ./Init.cc ./Log.cc ./pugixml/pugixml.cc ./qcd/action/fermion/CayleyFermion5D.cc ./qcd/action/fermion/ContinuedFractionFermion5D.cc ./qcd/action/fermion/PartialFractionFermion5D.cc ./qcd/action/fermion/WilsonFermion.cc ./qcd/action/fermion/WilsonFermion5D.cc ./qcd/action/fermion/WilsonKernels.cc ./qcd/action/fermion/WilsonKernelsHand.cc ./qcd/hmc/HMC.cc ./qcd/spin/Dirac.cc ./qcd/utils/SpaceTimeGrid.cc ./serialisation/BinaryIO.cc ./serialisation/TextIO.cc ./serialisation/XmlIO.cc ./stencil/Lebesgue.cc ./stencil/Stencil_common.cc
+CCFILES=./algorithms/approx/MultiShiftFunction.cc ./algorithms/approx/Remez.cc ./algorithms/approx/Zolotarev.cc ./Init.cc ./Log.cc ./PerfCount.cc ./pugixml/pugixml.cc ./qcd/action/fermion/CayleyFermion5D.cc ./qcd/action/fermion/ContinuedFractionFermion5D.cc ./qcd/action/fermion/PartialFractionFermion5D.cc ./qcd/action/fermion/WilsonFermion.cc ./qcd/action/fermion/WilsonFermion5D.cc ./qcd/action/fermion/WilsonKernels.cc ./qcd/action/fermion/WilsonKernelsAsm.cc ./qcd/action/fermion/WilsonKernelsHand.cc ./qcd/hmc/HMC.cc ./qcd/spin/Dirac.cc ./qcd/utils/SpaceTimeGrid.cc ./serialisation/BinaryIO.cc ./serialisation/TextIO.cc ./serialisation/XmlIO.cc ./stencil/Lebesgue.cc ./stencil/Stencil_common.cc
diff --git a/lib/PerfCount.cc b/lib/PerfCount.cc
index cc0464be..2b2476fa 100644
--- a/lib/PerfCount.cc
+++ b/lib/PerfCount.cc
@@ -3,8 +3,11 @@
 #include <PerfCount.h>
 
 namespace Grid {
+
 #define CacheControl(L,O,R) ((PERF_COUNT_HW_CACHE_##L)|(PERF_COUNT_HW_CACHE_OP_##O<<8)| (PERF_COUNT_HW_CACHE_RESULT_##R<<16))
+
 const PerformanceCounter::PerformanceCounterConfig PerformanceCounter::PerformanceCounterConfigs [] = {
+#ifdef __linux__
   { PERF_TYPE_HARDWARE, PERF_COUNT_HW_CPU_CYCLES          ,  "CPUCYCLES.........." },
   { PERF_TYPE_HARDWARE, PERF_COUNT_HW_INSTRUCTIONS        ,  "INSTRUCTIONS......." },
   { PERF_TYPE_HARDWARE, PERF_COUNT_HW_CACHE_REFERENCES    ,  "CACHE_REFERENCES..." },
@@ -23,6 +26,7 @@ const PerformanceCounter::PerformanceCounterConfig PerformanceCounter::Performan
   { PERF_TYPE_HW_CACHE, CacheControl(LL,PREFETCH,ACCESS)  ,  "LL_PREFETCH_ACCESS."},
   { PERF_TYPE_HW_CACHE, CacheControl(L1I,READ,MISS)       ,  "L1I_READ_MISS......"},
   { PERF_TYPE_HW_CACHE, CacheControl(L1I,READ,ACCESS)     ,  "L1I_READ_ACCESS...."}
+#endif
   //  { PERF_TYPE_HARDWARE, PERF_COUNT_HW_STALLED_CYCLES_FRONTEND, "STALL_CYCLES" },
 };
 }
diff --git a/lib/PerfCount.h b/lib/PerfCount.h
index c379639d..49a73fe6 100644
--- a/lib/PerfCount.h
+++ b/lib/PerfCount.h
@@ -30,6 +30,7 @@ static long perf_event_open(struct perf_event_attr *hw_event, pid_t pid,
 
 class PerformanceCounter {
 private:
+
   typedef struct { 
   public:
     uint32_t type;
diff --git a/lib/Simd.h b/lib/Simd.h
index 42d5e74c..3aaefddd 100644
--- a/lib/Simd.h
+++ b/lib/Simd.h
@@ -11,7 +11,6 @@
 // Vector types are arch dependent
 ////////////////////////////////////////////////////////////////////////
 
-typedef uint32_t Integer;
 
 #define _MM_SELECT_FOUR_FOUR(A,B,C,D) ((A<<6)|(B<<4)|(C<<2)|(D))
 #define _MM_SELECT_EIGHT_TWO(A,B,C,D,E,F,G,H) ((A<<7)|(B<<6)|(C<<5)|(D<<4)|(E<<3)|(F<<2)|(G<<4)|(H))
@@ -20,6 +19,8 @@ typedef uint32_t Integer;
 
 namespace Grid {
 
+  typedef uint32_t Integer;
+
   typedef  float  RealF;
   typedef  double RealD;
 #ifdef GRID_DEFAULT_PRECISION_DOUBLE
diff --git a/lib/qcd/action/fermion/CayleyFermion5D.cc b/lib/qcd/action/fermion/CayleyFermion5D.cc
index 39d04492..367e57dd 100644
--- a/lib/qcd/action/fermion/CayleyFermion5D.cc
+++ b/lib/qcd/action/fermion/CayleyFermion5D.cc
@@ -114,7 +114,7 @@ namespace QCD {
     // Apply Dw
     this->DW(psi,Din,DaggerYes); 
 
-    Meooe5D(Din,chi);
+    MeooeDag5D(Din,chi);
 
     int Ls=this->Ls;
     for(int s=0;s<Ls;s++){
@@ -163,7 +163,6 @@ namespace QCD {
     FermionField tmp(psi._grid);
     // Assemble the 5d matrix
     Meooe5D(psi,tmp); 
-
 #if 0
     std::cout << "Meooe Test replacement norm2 tmp = " <<norm2(tmp)<<std::endl;
     for(int s=0;s<Ls;s++){
@@ -202,7 +201,7 @@ namespace QCD {
       this->DhopOE(psi,tmp,DaggerYes);
     }
 
-    Meooe5D(tmp,chi); 
+    MeooeDag5D(tmp,chi); 
 #if 0
     std::cout << "Meooe Test replacement norm2 chi new = " <<norm2(chi)<<std::endl;
     // Assemble the 5d matrix
diff --git a/lib/qcd/action/fermion/WilsonKernels.cc b/lib/qcd/action/fermion/WilsonKernels.cc
index a37e7f9e..04d99c57 100644
--- a/lib/qcd/action/fermion/WilsonKernels.cc
+++ b/lib/qcd/action/fermion/WilsonKernels.cc
@@ -3,9 +3,12 @@ namespace Grid {
 namespace QCD {
 
 template<class Impl> 
-void WilsonKernels<Impl>::DiracOptDhopSite(StencilImpl &st,DoubledGaugeField &U,
-						  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
-						  int sF,int sU,const FermionField &in, FermionField &out)
+WilsonKernels<Impl>::WilsonKernels(const ImplParams &p): Base(p) {};
+
+template<class Impl> 
+void WilsonKernels<Impl>::DiracOptDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+					   std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+					   int sF,int sU,const FermionField &in, FermionField &out)
 {
   SiteHalfSpinor  tmp;    
   SiteHalfSpinor  chi;    
@@ -122,7 +125,7 @@ void WilsonKernels<Impl>::DiracOptDhopSite(StencilImpl &st,DoubledGaugeField &U,
 };
 
 template<class Impl> 
-void WilsonKernels<Impl>::DiracOptDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+void WilsonKernels<Impl>::DiracOptDhopSite(StencilImpl &st,DoubledGaugeField &U,
 					      std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 					      int sF,int sU,const FermionField &in, FermionField &out)
 {
@@ -369,6 +372,16 @@ void WilsonKernels<Impl>::DiracOptDhopDir(StencilImpl &st,DoubledGaugeField &U,
   vstream(out._odata[sF],result*(-0.5));
 }
 
+#if ( ! defined(AVX512) ) && ( ! defined(IMCI) )
+template<class Impl> 
+void WilsonKernels<Impl>::DiracOptAsmDhopSite(StencilImpl &st,DoubledGaugeField &U,
+					      std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+					      int sF,int sU,const FermionField &in, FermionField &out,uint64_t *p)
+{
+  DiracOptDhopSite(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
+}
+#endif
+
   FermOpTemplateInstantiate(WilsonKernels);
 
 }}
diff --git a/lib/qcd/action/fermion/WilsonKernels.h b/lib/qcd/action/fermion/WilsonKernels.h
index 988bc6f8..065ce678 100644
--- a/lib/qcd/action/fermion/WilsonKernels.h
+++ b/lib/qcd/action/fermion/WilsonKernels.h
@@ -28,19 +28,12 @@ namespace Grid {
      void DiracOptDhopDir(StencilImpl &st,DoubledGaugeField &U,
 			  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 			  int sF,int sU,const FermionField &in, FermionField &out,int dirdisp,int gamma);
-#if defined(AVX512) || defined(IMCI)
+
      void DiracOptAsmDhopSite(StencilImpl &st,DoubledGaugeField &U,
 			      std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 			      int sF,int sU,const FermionField &in, FermionField &out,uint64_t *);
-#else
-     void DiracOptAsmDhopSite(StencilImpl &st,DoubledGaugeField &U,
-			      std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
-			      int sF,int sU,const FermionField &in, FermionField &out,uint64_t *p){
-       DiracOptDhopSite(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
-     }
-#endif
+
 #define HANDOPT
-#ifdef HANDOPT
      void DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
 			       std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 			       int sF,int sU,const FermionField &in, FermionField &out);
@@ -48,25 +41,9 @@ namespace Grid {
      void DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
 				  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 				  int sF,int sU,const FermionField &in, FermionField &out);
-#else
-
-     void DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
-			       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
-     }
-
-     void DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
-				  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
-     }
-#endif
-
-     WilsonKernels(const ImplParams &p= ImplParams()) : Base(p) {};
 
+     WilsonKernels(const ImplParams &p= ImplParams());
+     
     };
 
   }
diff --git a/lib/qcd/action/fermion/WilsonKernelsHand.cc b/lib/qcd/action/fermion/WilsonKernelsHand.cc
index b263ea83..c1d01f51 100644
--- a/lib/qcd/action/fermion/WilsonKernelsHand.cc
+++ b/lib/qcd/action/fermion/WilsonKernelsHand.cc
@@ -282,7 +282,7 @@ namespace QCD {
 
 #ifdef HANDOPT
 template<class Impl>
-void WilsonKernels<Impl >::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
+void WilsonKernels<Impl >::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
 					       std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 					       int ss,int sU,const FermionField &in, FermionField &out)
 {
@@ -526,7 +526,7 @@ void WilsonKernels<Impl >::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeFiel
 }
 
 template<class Impl>
-void WilsonKernels<Impl >::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+void WilsonKernels<Impl >::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
 						   std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
 						   int ss,int sU,const FermionField &in, FermionField &out)
 {
@@ -767,6 +767,36 @@ void WilsonKernels<Impl >::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeF
     vstream(ref()(3)(2),result_32*(-0.5));
   }
 }
-  FermOpTemplateInstantiate(WilsonKernels);
+#else 
+template<class Impl>
+void WilsonKernels<Impl >::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
+					       std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+					       int ss,int sU,const FermionField &in, FermionField &out)
+{
+  DiracOptDhopSite(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
+}
+
+template<class Impl>
+void WilsonKernels<Impl >::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+						   std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+						   int ss,int sU,const FermionField &in, FermionField &out)
+{
+  DiracOptDhopSiteDag(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
+}
+
 #endif
+
+template void WilsonKernels<WilsonImplF>::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
+							  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+							  int ss,int sU,const FermionField &in, FermionField &out);
+template void WilsonKernels<WilsonImplD>::DiracOptHandDhopSite(StencilImpl &st,DoubledGaugeField &U,
+							  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+							  int ss,int sU,const FermionField &in, FermionField &out);
+template void WilsonKernels<WilsonImplF>::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+							  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+							  int ss,int sU,const FermionField &in, FermionField &out);
+template void WilsonKernels<WilsonImplD>::DiracOptHandDhopSiteDag(StencilImpl &st,DoubledGaugeField &U,
+							  std::vector<SiteHalfSpinor,alignedAllocator<SiteHalfSpinor> >  &buf,
+							  int ss,int sU,const FermionField &in, FermionField &out);
+
 }}
diff --git a/lib/simd/Grid_avx.h b/lib/simd/Grid_avx.h
index 45b8ad0c..e264639e 100644
--- a/lib/simd/Grid_avx.h
+++ b/lib/simd/Grid_avx.h
@@ -16,6 +16,7 @@
 #define _mm256_set_m128i(hi,lo) _mm256_insertf128_si256(_mm256_castsi128_si256(lo),(hi),1)
 #endif
 
+namespace Grid {
 namespace Optimization {
 
   template<class vtype>
@@ -467,7 +468,7 @@ namespace Optimization {
 
 //////////////////////////////////////////////////////////////////////////////////////
 // Here assign types 
-namespace Grid {
+
   typedef __m256  SIMD_Ftype; // Single precision type
   typedef __m256d SIMD_Dtype; // Double precision type
   typedef __m256i SIMD_Itype; // Integer type
@@ -488,8 +489,8 @@ namespace Grid {
   typedef Optimization::Vstore   VstoreSIMD;
   typedef Optimization::Vset     VsetSIMD;
   typedef Optimization::Vstream  VstreamSIMD;
-  template <typename S, typename T> using ReduceSIMD = Optimization::Reduce<S,T>;
 
+  template <typename S, typename T> using ReduceSIMD = Optimization::Reduce<S,T>;
 
   // Arithmetic operations
   typedef Optimization::Sum         SumSIMD;
diff --git a/scripts/filelist b/scripts/filelist
index 99db51b2..fd0e74ad 100755
--- a/scripts/filelist
+++ b/scripts/filelist
@@ -11,8 +11,6 @@ echo CCFILES=$CCFILES >> Make.inc
 
 cd ..
 
-
-
 cd tests
 
 echo> Make.inc
@@ -32,6 +30,26 @@ echo ${BNAME}_LDADD=-lGrid>> Make.inc
 echo >> Make.inc
 done
 
+cd qdpxx
+
+echo> Make.inc
+TESTS=`ls T*.cc`
+TESTLIST=`echo ${TESTS} | sed s/.cc//g `
+
+echo > Make.inc
+echo bin_PROGRAMS = ${TESTLIST} >> Make.inc
+echo >> Make.inc
+
+for f in $TESTS
+do
+BNAME=`basename $f .cc`
+echo >> Make.inc
+echo ${BNAME}_SOURCES=$f  >> Make.inc
+echo ${BNAME}_LDADD=-lGrid>> Make.inc
+echo >> Make.inc
+done
+
+cd ..
 cd ..
 
 
diff --git a/tests/Make.inc b/tests/Make.inc
index e0a17529..83092c4b 100644
--- a/tests/Make.inc
+++ b/tests/Make.inc
@@ -1,5 +1,5 @@
 
-bin_PROGRAMS = Test_GaugeAction Test_cayley_cg Test_cayley_coarsen_support Test_cayley_even_odd Test_cayley_ldop_cr Test_cf_coarsen_support Test_cf_cr_unprec Test_cheby Test_contfrac_cg Test_contfrac_even_odd Test_contfrac_force Test_cshift Test_cshift_red_black Test_dwf_cg_prec Test_dwf_cg_schur Test_dwf_cg_unprec Test_dwf_cr_unprec Test_dwf_even_odd Test_dwf_force Test_dwf_fpgcr Test_dwf_hdcr Test_gamma Test_hmc_EODWFRatio Test_hmc_EOWilsonFermionGauge Test_hmc_EOWilsonRatio Test_hmc_WilsonFermionGauge Test_hmc_WilsonGauge Test_hmc_WilsonRatio Test_lie_generators Test_main Test_multishift_sqrt Test_nersc_io Test_partfrac_force Test_quenched_update Test_remez Test_rhmc_EOWilson1p1 Test_rhmc_EOWilsonRatio Test_rhmc_Wilson1p1 Test_rhmc_WilsonRatio Test_rng Test_rng_fixed Test_serialisation Test_simd Test_stencil Test_wilson_cg_prec Test_wilson_cg_schur Test_wilson_cg_unprec Test_wilson_cr_unprec Test_wilson_even_odd Test_wilson_force Test_wilson_force_phiMdagMphi Test_wilson_force_phiMphi Test_synthetic_lanczos
+bin_PROGRAMS = Test_GaugeAction Test_cayley_cg Test_cayley_coarsen_support Test_cayley_even_odd Test_cayley_ldop_cr Test_cf_coarsen_support Test_cf_cr_unprec Test_cheby Test_contfrac_cg Test_contfrac_even_odd Test_contfrac_force Test_cshift Test_cshift_red_black Test_dwf_cg_prec Test_dwf_cg_schur Test_dwf_cg_unprec Test_dwf_cr_unprec Test_dwf_even_odd Test_dwf_force Test_dwf_fpgcr Test_dwf_hdcr Test_dwf_lanczos Test_gamma Test_hmc_EODWFRatio Test_hmc_EOWilsonFermionGauge Test_hmc_EOWilsonRatio Test_hmc_WilsonFermionGauge Test_hmc_WilsonGauge Test_hmc_WilsonRatio Test_lie_generators Test_main Test_multishift_sqrt Test_nersc_io Test_partfrac_force Test_quenched_update Test_remez Test_rhmc_EOWilson1p1 Test_rhmc_EOWilsonRatio Test_rhmc_Wilson1p1 Test_rhmc_WilsonRatio Test_rng Test_rng_fixed Test_serialisation Test_simd Test_stencil Test_synthetic_lanczos Test_wilson_cg_prec Test_wilson_cg_schur Test_wilson_cg_unprec Test_wilson_cr_unprec Test_wilson_even_odd Test_wilson_force Test_wilson_force_phiMdagMphi Test_wilson_force_phiMphi 
 
 
 Test_GaugeAction_SOURCES=Test_GaugeAction.cc
@@ -98,8 +98,8 @@ Test_gparity_SOURCES=Test_gparity.cc
 Test_gparity_LDADD=-lGrid
 
 
-#Test_gpwilson_even_odd_SOURCES=Test_gpwilson_even_odd.cc
-#Test_gpwilson_even_odd_LDADD=-lGrid
+Test_gpwilson_even_odd_SOURCES=Test_gpwilson_even_odd.cc
+Test_gpwilson_even_odd_LDADD=-lGrid
 
 
 Test_hmc_EODWFRatio_SOURCES=Test_hmc_EODWFRatio.cc
@@ -225,3 +225,7 @@ Test_wilson_force_phiMdagMphi_LDADD=-lGrid
 Test_wilson_force_phiMphi_SOURCES=Test_wilson_force_phiMphi.cc
 Test_wilson_force_phiMphi_LDADD=-lGrid
 
+
+Test_zmm_SOURCES=Test_zmm.cc
+Test_zmm_LDADD=-lGrid
+
diff --git a/tests/Makefile.am b/tests/Makefile.am
index 83385001..30d22d84 100644
--- a/tests/Makefile.am
+++ b/tests/Makefile.am
@@ -1,5 +1,12 @@
 # additional include paths necessary to compile the C++ library
+
+SUBDIRS = 
+if BUILD_CHROMA_REGRESSION
+  SUBDIRS+= qdpxx
+endif
+
 AM_CXXFLAGS = -I$(top_srcdir)/lib
 AM_LDFLAGS = -L$(top_builddir)/lib
 
+
 include Make.inc
diff --git a/tests/Test_cayley_cg.cc b/tests/Test_cayley_cg.cc
index 0f4955fd..09a44498 100644
--- a/tests/Test_cayley_cg.cc
+++ b/tests/Test_cayley_cg.cc
@@ -66,7 +66,8 @@ int main (int argc, char ** argv)
   GridParallelRNG          RNG5(FGrid);  RNG5.SeedFixedIntegers(seeds5);
   GridParallelRNG          RNG4(UGrid);  RNG4.SeedFixedIntegers(seeds4);
 
-  LatticeGaugeField Umu(UGrid); random(RNG4,Umu);
+  LatticeGaugeField Umu(UGrid);
+  SU3::HotConfiguration(RNG4,Umu);
   std::vector<LatticeColourMatrix> U(4,UGrid);
 
   RealD mass=0.1;
diff --git a/tests/Test_dwf_even_odd.cc b/tests/Test_dwf_even_odd.cc
index 96853b4c..687acf52 100644
--- a/tests/Test_dwf_even_odd.cc
+++ b/tests/Test_dwf_even_odd.cc
@@ -27,8 +27,8 @@ int main (int argc, char ** argv)
 
   const int Ls=8;
   GridCartesian         * UGrid   = SpaceTimeGrid::makeFourDimGrid(GridDefaultLatt(), GridDefaultSimd(Nd,vComplex::Nsimd()),GridDefaultMpi());
-  GridRedBlackCartesian * UrbGrid = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid);
   GridCartesian         * FGrid   = SpaceTimeGrid::makeFiveDimGrid(Ls,UGrid);
+  GridRedBlackCartesian * UrbGrid = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid);
   GridRedBlackCartesian * FrbGrid = SpaceTimeGrid::makeFiveDimRedBlackGrid(Ls,UGrid);
 
   std::vector<int> seeds4({1,2,3,4});