mirror of
https://github.com/paboyle/Grid.git
synced 2025-11-04 05:54:32 +00:00
Merge branch 'master' of https://github.com/paboyle/Grid
This commit is contained in:
paboyle
180
lib/Config.h.in
180
lib/Config.h.in
@@ -1,180 +0,0 @@
|
||||
/* lib/Config.h.in. Generated from configure.ac by autoheader. */
|
||||
|
||||
/* AVX Intrinsics */
|
||||
#undef AVX1
|
||||
|
||||
/* AVX2 Intrinsics */
|
||||
#undef AVX2
|
||||
|
||||
/* AVX512 Intrinsics for Knights Landing */
|
||||
#undef AVX512
|
||||
|
||||
/* AVX Intrinsics with FMA4 */
|
||||
#undef AVXFMA4
|
||||
|
||||
/* EMPTY_SIMD only for DEBUGGING */
|
||||
#undef EMPTY_SIMD
|
||||
|
||||
/* GRID_COMMS_MPI */
|
||||
#undef GRID_COMMS_MPI
|
||||
|
||||
/* GRID_COMMS_NONE */
|
||||
#undef GRID_COMMS_NONE
|
||||
|
||||
/* GRID_COMMS_SHMEM */
|
||||
#undef GRID_COMMS_SHMEM
|
||||
|
||||
/* GRID_DEFAULT_PRECISION is DOUBLE */
|
||||
#undef GRID_DEFAULT_PRECISION_DOUBLE
|
||||
|
||||
/* GRID_DEFAULT_PRECISION is SINGLE */
|
||||
#undef GRID_DEFAULT_PRECISION_SINGLE
|
||||
|
||||
/* Support Altivec instructions */
|
||||
#undef HAVE_ALTIVEC
|
||||
|
||||
/* Support AVX (Advanced Vector Extensions) instructions */
|
||||
#undef HAVE_AVX
|
||||
|
||||
/* Support AVX2 (Advanced Vector Extensions 2) instructions */
|
||||
#undef HAVE_AVX2
|
||||
|
||||
/* Define to 1 if you have the declaration of `be64toh', and to 0 if you
|
||||
don't. */
|
||||
#undef HAVE_DECL_BE64TOH
|
||||
|
||||
/* Define to 1 if you have the declaration of `ntohll', and to 0 if you don't.
|
||||
*/
|
||||
#undef HAVE_DECL_NTOHLL
|
||||
|
||||
/* Define to 1 if you have the <endian.h> header file. */
|
||||
#undef HAVE_ENDIAN_H
|
||||
|
||||
/* Define to 1 if you have the <execinfo.h> header file. */
|
||||
#undef HAVE_EXECINFO_H
|
||||
|
||||
/* Support FMA3 (Fused Multiply-Add) instructions */
|
||||
#undef HAVE_FMA
|
||||
|
||||
/* Define to 1 if you have the `gettimeofday' function. */
|
||||
#undef HAVE_GETTIMEOFDAY
|
||||
|
||||
/* Define to 1 if you have the <gmp.h> header file. */
|
||||
#undef HAVE_GMP_H
|
||||
|
||||
/* Define to 1 if you have the <inttypes.h> header file. */
|
||||
#undef HAVE_INTTYPES_H
|
||||
|
||||
/* Define to 1 if you have the <malloc.h> header file. */
|
||||
#undef HAVE_MALLOC_H
|
||||
|
||||
/* Define to 1 if you have the <malloc/malloc.h> header file. */
|
||||
#undef HAVE_MALLOC_MALLOC_H
|
||||
|
||||
/* Define to 1 if you have the <memory.h> header file. */
|
||||
#undef HAVE_MEMORY_H
|
||||
|
||||
/* Support mmx instructions */
|
||||
#undef HAVE_MMX
|
||||
|
||||
/* Define to 1 if you have the <mm_malloc.h> header file. */
|
||||
#undef HAVE_MM_MALLOC_H
|
||||
|
||||
/* Support SSE (Streaming SIMD Extensions) instructions */
|
||||
#undef HAVE_SSE
|
||||
|
||||
/* Support SSE2 (Streaming SIMD Extensions 2) instructions */
|
||||
#undef HAVE_SSE2
|
||||
|
||||
/* Support SSE3 (Streaming SIMD Extensions 3) instructions */
|
||||
#undef HAVE_SSE3
|
||||
|
||||
/* Support SSSE4.1 (Streaming SIMD Extensions 4.1) instructions */
|
||||
#undef HAVE_SSE4_1
|
||||
|
||||
/* Support SSSE4.2 (Streaming SIMD Extensions 4.2) instructions */
|
||||
#undef HAVE_SSE4_2
|
||||
|
||||
/* Support SSSE3 (Supplemental Streaming SIMD Extensions 3) instructions */
|
||||
#undef HAVE_SSSE3
|
||||
|
||||
/* Define to 1 if you have the <stdint.h> header file. */
|
||||
#undef HAVE_STDINT_H
|
||||
|
||||
/* Define to 1 if you have the <stdlib.h> header file. */
|
||||
#undef HAVE_STDLIB_H
|
||||
|
||||
/* Define to 1 if you have the <strings.h> header file. */
|
||||
#undef HAVE_STRINGS_H
|
||||
|
||||
/* Define to 1 if you have the <string.h> header file. */
|
||||
#undef HAVE_STRING_H
|
||||
|
||||
/* Define to 1 if you have the <sys/stat.h> header file. */
|
||||
#undef HAVE_SYS_STAT_H
|
||||
|
||||
/* Define to 1 if you have the <sys/types.h> header file. */
|
||||
#undef HAVE_SYS_TYPES_H
|
||||
|
||||
/* Define to 1 if you have the <unistd.h> header file. */
|
||||
#undef HAVE_UNISTD_H
|
||||
|
||||
/* IMCI Intrinsics for Knights Corner */
|
||||
#undef IMCI
|
||||
|
||||
/* NEON ARMv8 Experimental support */
|
||||
#undef NEONv8
|
||||
|
||||
/* Name of package */
|
||||
#undef PACKAGE
|
||||
|
||||
/* Define to the address where bug reports for this package should be sent. */
|
||||
#undef PACKAGE_BUGREPORT
|
||||
|
||||
/* Define to the full name of this package. */
|
||||
#undef PACKAGE_NAME
|
||||
|
||||
/* Define to the full name and version of this package. */
|
||||
#undef PACKAGE_STRING
|
||||
|
||||
/* Define to the one symbol short name of this package. */
|
||||
#undef PACKAGE_TARNAME
|
||||
|
||||
/* Define to the version of this package. */
|
||||
#undef PACKAGE_VERSION
|
||||
|
||||
/* RNG_MT19937 */
|
||||
#undef RNG_MT19937
|
||||
|
||||
/* RNG_RANLUX */
|
||||
#undef RNG_RANLUX
|
||||
|
||||
/* SSE4 Intrinsics */
|
||||
#undef SSE4
|
||||
|
||||
/* Define to 1 if you have the ANSI C header files. */
|
||||
#undef STDC_HEADERS
|
||||
|
||||
/* Version number of package */
|
||||
#undef VERSION
|
||||
|
||||
/* Define for Solaris 2.5.1 so the uint32_t typedef from <sys/synch.h>,
|
||||
<pthread.h>, or <semaphore.h> is not used. If the typedef were allowed, the
|
||||
#define below would cause a syntax error. */
|
||||
#undef _UINT32_T
|
||||
|
||||
/* Define for Solaris 2.5.1 so the uint64_t typedef from <sys/synch.h>,
|
||||
<pthread.h>, or <semaphore.h> is not used. If the typedef were allowed, the
|
||||
#define below would cause a syntax error. */
|
||||
#undef _UINT64_T
|
||||
|
||||
/* Define to `unsigned int' if <sys/types.h> does not define. */
|
||||
#undef size_t
|
||||
|
||||
/* Define to the type of an unsigned integer type of width exactly 32 bits if
|
||||
such a type exists and the standard includes do not define it. */
|
||||
#undef uint32_t
|
||||
|
||||
/* Define to the type of an unsigned integer type of width exactly 64 bits if
|
||||
such a type exists and the standard includes do not define it. */
|
||||
#undef uint64_t
|
||||
@@ -152,16 +152,18 @@ namespace Grid {
|
||||
{
|
||||
// Run a Lanczos with sloppy convergence
|
||||
const int Nstop = nn;
|
||||
const int Nk = nn+10;
|
||||
const int Np = nn+10;
|
||||
const int Nk = nn+20;
|
||||
const int Np = nn+20;
|
||||
const int Nm = Nk+Np;
|
||||
const int MaxIt= 10000;
|
||||
RealD resid = 1.0e-5;
|
||||
RealD resid = 1.0e-3;
|
||||
|
||||
Chebyshev<FineField> Cheb(0.2,5.,11);
|
||||
Chebyshev<FineField> Cheb(0.5,64.0,21);
|
||||
ImplicitlyRestartedLanczos<FineField> IRL(hermop,Cheb,Nstop,Nk,Nm,resid,MaxIt);
|
||||
// IRL.lock = 1;
|
||||
|
||||
FineField noise(FineGrid); gaussian(RNG,noise);
|
||||
FineField tmp(FineGrid);
|
||||
std::vector<RealD> eval(Nm);
|
||||
std::vector<FineField> evec(Nm,FineGrid);
|
||||
|
||||
@@ -172,16 +174,34 @@ namespace Grid {
|
||||
|
||||
// pull back nn vectors
|
||||
for(int b=0;b<nn;b++){
|
||||
|
||||
subspace[b] = evec[b];
|
||||
|
||||
std::cout << GridLogMessage <<"subspace["<<b<<"] = "<<norm2(subspace[b])<<std::endl;
|
||||
|
||||
hermop.Op(subspace[b],tmp);
|
||||
std::cout<<GridLogMessage << "filtered["<<b<<"] <f|MdagM|f> "<<norm2(tmp)<<std::endl;
|
||||
|
||||
noise = tmp - sqrt(eval[b])*subspace[b] ;
|
||||
|
||||
std::cout<<GridLogMessage << " lambda_"<<b<<" = "<< eval[b] <<" ; [ M - Lambda ]_"<<b<<" vec_"<<b<<" = " <<norm2(noise)<<std::endl;
|
||||
|
||||
noise = tmp + eval[b]*subspace[b] ;
|
||||
|
||||
std::cout<<GridLogMessage << " lambda_"<<b<<" = "<< eval[b] <<" ; [ M - Lambda ]_"<<b<<" vec_"<<b<<" = " <<norm2(noise)<<std::endl;
|
||||
|
||||
}
|
||||
Orthogonalise();
|
||||
for(int b=0;b<nn;b++){
|
||||
std::cout << GridLogMessage <<"subspace["<<b<<"] = "<<norm2(subspace[b])<<std::endl;
|
||||
}
|
||||
}
|
||||
|
||||
virtual void CreateSubspace(GridParallelRNG &RNG,LinearOperatorBase<FineField> &hermop,int nn=nbasis) {
|
||||
|
||||
RealD scale;
|
||||
|
||||
ConjugateGradient<FineField> CG(5.0e-3,10000);
|
||||
ConjugateGradient<FineField> CG(1.0e-2,10000);
|
||||
FineField noise(FineGrid);
|
||||
FineField Mn(FineGrid);
|
||||
|
||||
|
||||
@@ -39,42 +39,33 @@ class SortEigen {
|
||||
private:
|
||||
|
||||
//hacking for testing for now
|
||||
#if 0
|
||||
static bool less_lmd(RealD left,RealD right){
|
||||
return fabs(left) < fabs(right);
|
||||
}
|
||||
static bool less_pair(std::pair<RealD,Field>& left,
|
||||
std::pair<RealD,Field>& right){
|
||||
return fabs(left.first) < fabs(right.first);
|
||||
}
|
||||
#else
|
||||
private:
|
||||
static bool less_lmd(RealD left,RealD right){
|
||||
return left > right;
|
||||
}
|
||||
static bool less_pair(std::pair<RealD,Field>& left,
|
||||
std::pair<RealD,Field>& right){
|
||||
static bool less_pair(std::pair<RealD,Field const*>& left,
|
||||
std::pair<RealD,Field const*>& right){
|
||||
return left.first > (right.first);
|
||||
}
|
||||
#endif
|
||||
|
||||
|
||||
public:
|
||||
|
||||
void push(DenseVector<RealD>& lmd,
|
||||
DenseVector<Field>& evec,int N) {
|
||||
DenseVector<Field>& evec,int N) {
|
||||
DenseVector<Field> cpy(lmd.size(),evec[0]._grid);
|
||||
for(int i=0;i<lmd.size();i++) cpy[i] = evec[i];
|
||||
|
||||
DenseVector<std::pair<RealD, Field> > emod;
|
||||
typename DenseVector<std::pair<RealD, Field> >::iterator it;
|
||||
|
||||
for(int i=0;i<lmd.size();++i){
|
||||
emod.push_back(std::pair<RealD,Field>(lmd[i],evec[i]));
|
||||
}
|
||||
DenseVector<std::pair<RealD, Field const*> > emod(lmd.size());
|
||||
for(int i=0;i<lmd.size();++i)
|
||||
emod[i] = std::pair<RealD,Field const*>(lmd[i],&cpy[i]);
|
||||
|
||||
partial_sort(emod.begin(),emod.begin()+N,emod.end(),less_pair);
|
||||
|
||||
it=emod.begin();
|
||||
typename DenseVector<std::pair<RealD, Field const*> >::iterator it = emod.begin();
|
||||
for(int i=0;i<N;++i){
|
||||
lmd[i]=it->first;
|
||||
evec[i]=it->second;
|
||||
evec[i]=*(it->second);
|
||||
++it;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -637,21 +637,20 @@ until convergence
|
||||
abort();
|
||||
|
||||
converged:
|
||||
// Sorting
|
||||
// Sorting
|
||||
eval.resize(Nconv);
|
||||
evec.resize(Nconv,grid);
|
||||
for(int i=0; i<Nconv; ++i){
|
||||
eval[i] = eval2[Iconv[i]];
|
||||
evec[i] = B[Iconv[i]];
|
||||
}
|
||||
_sort.push(eval,evec,Nconv);
|
||||
|
||||
eval.clear();
|
||||
evec.clear();
|
||||
for(int i=0; i<Nconv; ++i){
|
||||
eval.push_back(eval2[Iconv[i]]);
|
||||
evec.push_back(B[Iconv[i]]);
|
||||
}
|
||||
_sort.push(eval,evec,Nconv);
|
||||
|
||||
std::cout << "\n Converged\n Summary :\n";
|
||||
std::cout << " -- Iterations = "<< Nconv << "\n";
|
||||
std::cout << " -- beta(k) = "<< beta_k << "\n";
|
||||
std::cout << " -- Nconv = "<< Nconv << "\n";
|
||||
}
|
||||
std::cout << "\n Converged\n Summary :\n";
|
||||
std::cout << " -- Iterations = "<< Nconv << "\n";
|
||||
std::cout << " -- beta(k) = "<< beta_k << "\n";
|
||||
std::cout << " -- Nconv = "<< Nconv << "\n";
|
||||
}
|
||||
|
||||
/////////////////////////////////////////////////
|
||||
// Adapted from Rudy's lanczos factor routine
|
||||
|
||||
@@ -522,7 +522,7 @@ void WilsonKernels<Impl>::DiracOptDhopDir(StencilImpl &st,DoubledGaugeField &U,
|
||||
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,bool local, bool nonlocal)
|
||||
int sF,int sU,const FermionField &in, FermionField &out)
|
||||
{
|
||||
DiracOptDhopSite(st,U,buf,sF,sU,in,out); // will template override for Wilson Nc=3
|
||||
}
|
||||
|
||||
@@ -329,7 +329,7 @@ public:
|
||||
CoarseVector Ctmp(_CoarseOperator.Grid());
|
||||
CoarseVector Csol(_CoarseOperator.Grid()); Csol=zero;
|
||||
|
||||
ConjugateGradient<CoarseVector> CG(1.0e-4,100000);
|
||||
ConjugateGradient<CoarseVector> CG(3.0e-3,100000);
|
||||
// ConjugateGradient<FineField> fCG(3.0e-2,1000);
|
||||
|
||||
HermitianLinearOperator<CoarseOperator,CoarseVector> HermOp(_CoarseOperator);
|
||||
@@ -560,8 +560,8 @@ int main (int argc, char ** argv)
|
||||
assert ( (nbasis & 0x1)==0);
|
||||
int nb=nbasis/2;
|
||||
std::cout<<GridLogMessage << " nbasis/2 = "<<nb<<std::endl;
|
||||
Aggregates.CreateSubspace(RNG5,HermDefOp,nb);
|
||||
// Aggregates.CreateSubspaceLanczos(RNG5,HermDefOp,nb);
|
||||
// Aggregates.CreateSubspace(RNG5,HermDefOp,nb);
|
||||
Aggregates.CreateSubspaceLanczos(RNG5,HermDefOp,nb);
|
||||
for(int n=0;n<nb;n++){
|
||||
G5R5(Aggregates.subspace[n+nb],Aggregates.subspace[n]);
|
||||
std::cout<<GridLogMessage<<n<<" subspace "<<norm2(Aggregates.subspace[n+nb])<<" "<<norm2(Aggregates.subspace[n]) <<std::endl;
|
||||
|
||||
Reference in New Issue
Block a user