Schur complement based red-black inversion working

2024-11-09 23:45:36 +00:00 · 2015-05-25 13:47:12 +01:00 · 2015-05-25 13:47:12 +01:00 · 489b1b9633
commit 489b1b9633
parent ea3240ad55
8 changed files with 336 additions and 34 deletions
--- a/benchmarks/Grid_wilson.cc
+++ b/benchmarks/Grid_wilson.cc
@ -24,7 +24,8 @@ int main (int argc, char ** argv)
  std::vector<int> latt_size   = GridDefaultLatt();
  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplexF::Nsimd());
  std::vector<int> mpi_layout  = GridDefaultMpi();
-  GridCartesian     Grid(latt_size,simd_layout,mpi_layout);
+  GridCartesian               Grid(latt_size,simd_layout,mpi_layout);
+  GridRedBlackCartesian     RBGrid(latt_size,simd_layout,mpi_layout);

  int threads = GridThread::GetThreads();
  std::cout << "Grid is setup to use "<<threads<<" threads"<<std::endl;
@ -36,11 +37,11 @@ int main (int argc, char ** argv)
  //  pRNG.SeedFixedIntegers(seeds);
  pRNG.SeedRandomDevice();

-  LatticeFermion src(&Grid); random(pRNG,src);
+  LatticeFermion src   (&Grid); random(pRNG,src);
  LatticeFermion result(&Grid); result=zero;
  LatticeFermion    ref(&Grid);    ref=zero;
-  LatticeFermion    err(&Grid);    
  LatticeFermion    tmp(&Grid);    tmp=zero;
+  LatticeFermion    err(&Grid);    tmp=zero;
  LatticeGaugeField Umu(&Grid); random(pRNG,Umu);
  std::vector<LatticeColourMatrix> U(4,&Grid);

@ -51,8 +52,11 @@ int main (int argc, char ** argv)

  // Only one non-zero (y)
  Umu=zero;
+  Complex cone(1.0,0.0);
  for(int nn=0;nn<Nd;nn++){
    random(pRNG,U[nn]);
+    if (nn!=0) U[nn]=zero;
+    else U[nn] = cone;
    pokeIndex<LorentzIndex>(Umu,U[nn],nn);
  }

@ -78,7 +82,7 @@ int main (int argc, char ** argv)
  }

  RealD mass=0.1;
-  WilsonMatrix Dw(Umu,mass);
+  WilsonMatrix Dw(Umu,Grid,RBGrid,mass);
  
  std::cout << "Calling Dw"<<std::endl;
  int ncall=1000;
@ -93,7 +97,7 @@ int main (int argc, char ** argv)
  std::cout << "norm result "<< norm2(result)<<std::endl;
  std::cout << "norm ref    "<< norm2(ref)<<std::endl;
  std::cout << "mflop/s =   "<< flops/(t1-t0)<<std::endl;
-  err = ref -result;
+  err = ref-result; 
  std::cout << "norm diff   "<< norm2(err)<<std::endl;


@ -129,9 +133,8 @@ int main (int argc, char ** argv)
  std::cout << "Called DwDag"<<std::endl;
  std::cout << "norm result "<< norm2(result)<<std::endl;
  std::cout << "norm ref    "<< norm2(ref)<<std::endl;
-  err = ref -result;
+  err = ref-result; 
  std::cout << "norm diff   "<< norm2(err)<<std::endl;

-
  Grid_finalize();
 }
--- a/benchmarks/Grid_wilson_cg_prec.cc
+++ b/benchmarks/Grid_wilson_cg_prec.cc
@ -0,0 +1,61 @@
+#include <Grid.h>
+
+using namespace std;
+using namespace Grid;
+using namespace Grid::QCD;
+
+template<class d>
+struct scal {
+  d internal;
+};
+
+  Gamma::GammaMatrix Gmu [] = {
+    Gamma::GammaX,
+    Gamma::GammaY,
+    Gamma::GammaZ,
+    Gamma::GammaT
+  };
+
+int main (int argc, char ** argv)
+{
+  Grid_init(&argc,&argv);
+
+
+  std::vector<int> latt_size   = GridDefaultLatt();
+  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplexF::Nsimd());
+  std::vector<int> mpi_layout  = GridDefaultMpi();
+  GridCartesian               Grid(latt_size,simd_layout,mpi_layout);
+  GridRedBlackCartesian     RBGrid(latt_size,simd_layout,mpi_layout);
+
+  std::vector<int> seeds({1,2,3,4});
+  GridParallelRNG          pRNG(&Grid);  pRNG.SeedFixedIntegers(seeds);
+
+  LatticeFermion src(&Grid); random(pRNG,src);
+  RealD nrm = norm2(src);
+  LatticeFermion result(&Grid); result=zero;
+  LatticeGaugeField Umu(&Grid); random(pRNG,Umu);
+
+  std::vector<LatticeColourMatrix> U(4,&Grid);
+
+  for(int mu=0;mu<Nd;mu++){
+    U[mu] = peekIndex<LorentzIndex>(Umu,mu);
+  }
+  
+  RealD mass=0.5;
+  WilsonMatrix Dw(Umu,Grid,RBGrid,mass);
+
+  //  HermitianOperator<WilsonMatrix,LatticeFermion> HermOp(Dw);
+  //  ConjugateGradient<LatticeFermion> CG(1.0e-8,10000);
+  //  CG(HermOp,src,result);
+  
+  LatticeFermion    src_o(&RBGrid);
+  LatticeFermion result_o(&RBGrid);
+  pickCheckerboard(Odd,src_o,src);
+  result_o=zero;
+
+  HermitianCheckerBoardedOperator<WilsonMatrix,LatticeFermion> HermOpEO(Dw);
+  ConjugateGradient<LatticeFermion> CG(1.0e-8,10000);
+  CG(HermOpEO,src_o,result_o);
+
+  Grid_finalize();
+}
--- a/benchmarks/Grid_wilson_cg_schur.cc
+++ b/benchmarks/Grid_wilson_cg_schur.cc
@ -0,0 +1,48 @@
+#include <Grid.h>
+
+using namespace std;
+using namespace Grid;
+using namespace Grid::QCD;
+
+template<class d>
+struct scal {
+  d internal;
+};
+
+  Gamma::GammaMatrix Gmu [] = {
+    Gamma::GammaX,
+    Gamma::GammaY,
+    Gamma::GammaZ,
+    Gamma::GammaT
+  };
+
+int main (int argc, char ** argv)
+{
+  Grid_init(&argc,&argv);
+
+
+  std::vector<int> latt_size   = GridDefaultLatt();
+  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplexF::Nsimd());
+  std::vector<int> mpi_layout  = GridDefaultMpi();
+  GridCartesian               Grid(latt_size,simd_layout,mpi_layout);
+  GridRedBlackCartesian     RBGrid(latt_size,simd_layout,mpi_layout);
+
+  std::vector<int> seeds({1,2,3,4});
+  GridParallelRNG          pRNG(&Grid);  pRNG.SeedFixedIntegers(seeds);
+
+  LatticeGaugeField Umu(&Grid); random(pRNG,Umu);
+
+  LatticeFermion    src(&Grid); random(pRNG,src);
+  LatticeFermion result(&Grid); result=zero;
+  LatticeFermion resid(&Grid); 
+
+  RealD mass=0.5;
+  WilsonMatrix Dw(Umu,Grid,RBGrid,mass);
+
+  ConjugateGradient<LatticeFermion> CG(1.0e-8,10000);
+  SchurRedBlackSolve<LatticeFermion> SchurSolver(CG);
+
+  SchurSolver(Dw,src,result);
+  
+  Grid_finalize();
+}
--- a/benchmarks/Grid_wilson_cg_unprec.cc
+++ b/benchmarks/Grid_wilson_cg_unprec.cc
@ -24,7 +24,8 @@ int main (int argc, char ** argv)
  std::vector<int> latt_size   = GridDefaultLatt();
  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplexF::Nsimd());
  std::vector<int> mpi_layout  = GridDefaultMpi();
-  GridCartesian     Grid(latt_size,simd_layout,mpi_layout);
+  GridCartesian               Grid(latt_size,simd_layout,mpi_layout);
+  GridRedBlackCartesian     RBGrid(latt_size,simd_layout,mpi_layout);

  std::vector<int> seeds({1,2,3,4});
  GridParallelRNG          pRNG(&Grid);  pRNG.SeedFixedIntegers(seeds);
@ -46,11 +47,11 @@ int main (int argc, char ** argv)
  }
  
  RealD mass=0.5;
-  WilsonMatrix Dw(Umu,mass);
+  WilsonMatrix Dw(Umu,Grid,RBGrid,mass);
+
  HermitianOperator<WilsonMatrix,LatticeFermion> HermOp(Dw);
  ConjugateGradient<LatticeFermion> CG(1.0e-8,10000);
  CG(HermOp,src,result);

-  
  Grid_finalize();
 }
--- a/benchmarks/Grid_wilson_evenodd.cc
+++ b/benchmarks/Grid_wilson_evenodd.cc
@ -0,0 +1,201 @@
+#include <Grid.h>
+
+using namespace std;
+using namespace Grid;
+using namespace Grid::QCD;
+
+template<class d>
+struct scal {
+  d internal;
+};
+
+  Gamma::GammaMatrix Gmu [] = {
+    Gamma::GammaX,
+    Gamma::GammaY,
+    Gamma::GammaZ,
+    Gamma::GammaT
+  };
+
+int main (int argc, char ** argv)
+{
+  Grid_init(&argc,&argv);
+
+  std::vector<int> latt_size   = GridDefaultLatt();
+  std::vector<int> simd_layout = GridDefaultSimd(Nd,vComplexF::Nsimd());
+  std::vector<int> mpi_layout  = GridDefaultMpi();
+  GridCartesian               Grid(latt_size,simd_layout,mpi_layout);
+  GridRedBlackCartesian     RBGrid(latt_size,simd_layout,mpi_layout);
+
+  int threads = GridThread::GetThreads();
+  std::cout << "Grid is setup to use "<<threads<<" threads"<<std::endl;
+
+  std::vector<int> seeds({1,2,3,4});
+
+  GridParallelRNG          pRNG(&Grid);
+  //  std::vector<int> seeds({1,2,3,4});
+  //  pRNG.SeedFixedIntegers(seeds);
+  pRNG.SeedRandomDevice();
+
+  LatticeFermion src   (&Grid); random(pRNG,src);
+  LatticeFermion phi   (&Grid); random(pRNG,phi);
+  LatticeFermion chi   (&Grid); random(pRNG,chi);
+  LatticeFermion result(&Grid); result=zero;
+  LatticeFermion    ref(&Grid);    ref=zero;
+  LatticeFermion    tmp(&Grid);    tmp=zero;
+  LatticeFermion    err(&Grid);    tmp=zero;
+  LatticeGaugeField Umu(&Grid); random(pRNG,Umu);
+  std::vector<LatticeColourMatrix> U(4,&Grid);
+
+  double volume=1;
+  for(int mu=0;mu<Nd;mu++){
+    volume=volume*latt_size[mu];
+  }  
+
+  // Only one non-zero (y)
+  Umu=zero;
+  for(int nn=0;nn<Nd;nn++){
+    random(pRNG,U[nn]);
+    pokeIndex<LorentzIndex>(Umu,U[nn],nn);
+  }
+
+  RealD mass=0.1;
+
+  WilsonMatrix Dw(Umu,Grid,RBGrid,mass);
+
+  LatticeFermion src_e   (&RBGrid);
+  LatticeFermion src_o   (&RBGrid);
+  LatticeFermion r_e   (&RBGrid);
+  LatticeFermion r_o   (&RBGrid);
+  LatticeFermion r_eo  (&Grid);
+
+  const int Even=0;
+  const int Odd=1;
+  std::cout<<"=========================================================="<<std::endl;
+  std::cout<<"= Testing that Deo + Doe = Dunprec "<<std::endl;
+  std::cout<<"=========================================================="<<std::endl;
+
+  pickCheckerboard(Even,src_e,src);
+  pickCheckerboard(Odd,src_o,src);
+
+  Dw.Meooe(src_e,r_o);  std::cout<<"Applied Meo"<<std::endl;
+  Dw.Meooe(src_o,r_e);  std::cout<<"Applied Moe"<<std::endl;
+  Dw.Dhop (src,ref,0);
+
+  setCheckerboard(r_eo,r_o);
+  setCheckerboard(r_eo,r_e);
+
+  ref = (-0.5)*ref;
+  err= ref - r_eo;
+  std::cout << "EO norm diff   "<< norm2(err)<< " "<<norm2(ref)<< " " << norm2(r_eo) <<std::endl;
+
+  LatticeComplex cerr(&Grid);
+  cerr = localInnerProduct(err,err);
+
+  std::cout<<"=============================================================="<<std::endl;
+  std::cout<<"= Test Ddagger is the dagger of D by requiring                "<<std::endl;
+  std::cout<<"=  < phi | Deo | chi > * = < chi | Deo^dag| phi>  "<<std::endl;
+  std::cout<<"=============================================================="<<std::endl;
+  
+  LatticeFermion chi_e   (&RBGrid);
+  LatticeFermion chi_o   (&RBGrid);
+
+  LatticeFermion dchi_e  (&RBGrid);
+  LatticeFermion dchi_o  (&RBGrid);
+
+  LatticeFermion phi_e   (&RBGrid);
+  LatticeFermion phi_o   (&RBGrid);
+
+  LatticeFermion dphi_e  (&RBGrid);
+  LatticeFermion dphi_o  (&RBGrid);
+
+
+  pickCheckerboard(Even,chi_e,chi);
+  pickCheckerboard(Odd ,chi_o,chi);
+  pickCheckerboard(Even,phi_e,phi);
+  pickCheckerboard(Odd ,phi_o,phi);
+
+  Dw.Meooe(chi_e,dchi_o);
+  Dw.Meooe(chi_o,dchi_e);
+  Dw.MeooeDag(phi_e,dphi_o);
+  Dw.MeooeDag(phi_o,dphi_e);
+
+  ComplexD pDce = innerProduct(phi_e,dchi_e);
+  ComplexD pDco = innerProduct(phi_o,dchi_o);
+  ComplexD cDpe = innerProduct(chi_e,dphi_e);
+  ComplexD cDpo = innerProduct(chi_o,dphi_o);
+
+  std::cout <<"e "<<pDce<<" "<<cDpe <<std::endl;
+  std::cout <<"o "<<pDco<<" "<<cDpo <<std::endl;
+
+  std::cout <<"pDce - conj(cDpo) "<< pDce-conj(cDpo) <<std::endl;
+  std::cout <<"pDco - conj(cDpe) "<< pDco-conj(cDpe) <<std::endl;
+
+  std::cout<<"=============================================================="<<std::endl;
+  std::cout<<"= Test MeeInv Mee = 1                                         "<<std::endl;
+  std::cout<<"=============================================================="<<std::endl;
+
+  pickCheckerboard(Even,chi_e,chi);
+  pickCheckerboard(Odd ,chi_o,chi);
+
+  Dw.Mooee(chi_e,src_e);
+  Dw.MooeeInv(src_e,phi_e);
+
+  Dw.Mooee(chi_o,src_o);
+  Dw.MooeeInv(src_o,phi_o);
+  
+  setCheckerboard(phi,phi_e);
+  setCheckerboard(phi,phi_o);
+
+  err = phi-chi;
+  std::cout << "norm diff   "<< norm2(err)<< std::endl;
+
+  std::cout<<"=============================================================="<<std::endl;
+  std::cout<<"= Test MeeInvDag MeeDag = 1                                   "<<std::endl;
+  std::cout<<"=============================================================="<<std::endl;
+
+  pickCheckerboard(Even,chi_e,chi);
+  pickCheckerboard(Odd ,chi_o,chi);
+
+  Dw.MooeeDag(chi_e,src_e);
+  Dw.MooeeInvDag(src_e,phi_e);
+
+  Dw.MooeeDag(chi_o,src_o);
+  Dw.MooeeInvDag(src_o,phi_o);
+  
+  setCheckerboard(phi,phi_e);
+  setCheckerboard(phi,phi_o);
+
+  err = phi-chi;
+  std::cout << "norm diff   "<< norm2(err)<< std::endl;
+
+  std::cout<<"=============================================================="<<std::endl;
+  std::cout<<"= Test MpcDagMpc is Hermitian              "<<std::endl;
+  std::cout<<"=============================================================="<<std::endl;
+  
+  random(pRNG,phi);
+  random(pRNG,chi);
+  pickCheckerboard(Even,chi_e,chi);
+  pickCheckerboard(Odd ,chi_o,chi);
+  pickCheckerboard(Even,phi_e,phi);
+  pickCheckerboard(Odd ,phi_o,phi);
+  RealD t1,t2;
+
+  Dw.MpcDagMpc(chi_e,dchi_e,t1,t2);
+  Dw.MpcDagMpc(chi_o,dchi_o,t1,t2);
+
+  Dw.MpcDagMpc(phi_e,dphi_e,t1,t2);
+  Dw.MpcDagMpc(phi_o,dphi_o,t1,t2);
+
+  pDce = innerProduct(phi_e,dchi_e);
+  pDco = innerProduct(phi_o,dchi_o);
+  cDpe = innerProduct(chi_e,dphi_e);
+  cDpo = innerProduct(chi_o,dphi_o);
+
+  std::cout <<"e "<<pDce<<" "<<cDpe <<std::endl;
+  std::cout <<"o "<<pDco<<" "<<cDpo <<std::endl;
+
+  std::cout <<"pDce - conj(cDpo) "<< pDco-conj(cDpo) <<std::endl;
+  std::cout <<"pDco - conj(cDpe) "<< pDce-conj(cDpe) <<std::endl;
+  
+  Grid_finalize();
+}
--- a/lib/simd/Grid_sse4.h
+++ b/lib/simd/Grid_sse4.h
@ -283,6 +283,7 @@ namespace Optimization {
 //////////////////////////////////////////////////////////////////////////////////////
 // Here assign types 
 namespace Grid {
+
  typedef __m128 SIMD_Ftype;  // Single precision type
  typedef __m128d SIMD_Dtype; // Double precision type
  typedef __m128i SIMD_Itype; // Integer type
--- a/lib/simd/Grid_vector_types.h
+++ b/lib/simd/Grid_vector_types.h
@ -13,31 +13,24 @@
 namespace Grid {

  // To take the floating point type of real/complex type
-  template <typename T> 
-    struct RealPart {
-      typedef T type;
-    };
-  template <typename T> 
-    struct RealPart< std::complex<T> >{
+  template <typename T> struct RealPart {
+    typedef T type;
+  };
+  template <typename T> struct RealPart< std::complex<T> >{
    typedef T type;
  };

  // type alias used to simplify the syntax of std::enable_if
-  template <typename T> using Invoke =
-    typename T::type;
-  template <typename Condition, typename ReturnType> using EnableIf =
-    Invoke<std::enable_if<Condition::value, ReturnType>>;
-  template <typename Condition, typename ReturnType> using NotEnableIf =
-    Invoke<std::enable_if<!Condition::value, ReturnType>>;
+  template <typename T> using Invoke                                  =  typename T::type;
+  template <typename Condition, typename ReturnType> using EnableIf   =    Invoke<std::enable_if<Condition::value, ReturnType>>;
+  template <typename Condition, typename ReturnType> using NotEnableIf=    Invoke<std::enable_if<!Condition::value, ReturnType>>;
  


  ////////////////////////////////////////////////////////
  // Check for complexity with type traits
-  template <typename T> 
-    struct is_complex : std::false_type {};
-  template < typename T > 
-    struct is_complex< std::complex<T> >: std::true_type {};
+  template <typename T>     struct is_complex : std::false_type {};
+  template < typename T >   struct is_complex< std::complex<T> >: std::true_type {};
  ////////////////////////////////////////////////////////
  // Define the operation templates functors
  // general forms to allow for vsplat syntax
@ -86,8 +79,6 @@ namespace Grid {
    Grid_simd(Real a){
      vsplat(*this,Scalar_type(a));
    };
-
-
       
    ///////////////////////////////////////////////
    // mac, mult, sub, add, adj
@ -126,10 +117,6 @@ namespace Grid {
      friend inline void vtrue (Grid_simd &ret){vsplat(ret,0xFFFFFFFF);}
    template <  class S = Scalar_type, EnableIf<std::is_integral < S >, int> = 0 > 
      friend inline void vfalse(vInteger &ret){vsplat(ret,0);}
-
-
-   
- 
   
    ////////////////////////////////////
    // Arithmetic operator overloads +,-,*
@ -165,7 +152,6 @@ namespace Grid {
 	ret.v = binary<Vector_type>(a.v,b.v, MultSIMD());
 	return ret;
      };
-    

    ////////////////////////////////////////////////////////////////////////
    // FIXME:  gonna remove these load/store, get, set, prefetch
--- a/tests/Grid_main.cc
+++ b/tests/Grid_main.cc
@ -1,8 +1,9 @@
 #include "Grid.h"

 //DEBUG
+#ifdef SSE4
 #include "simd/Grid_vector_types.h"
-
+#endif

 using namespace std;
 using namespace Grid;