Inverters
[Fermion action support]


Namespaces
namespace	Chroma::InvEigCG2Env
	Conjugate-Gradient (CGNE) with eigenvector acceleration.
namespace	Chroma::InvEigCG2ArrayEnv
	Conjugate-Gradient (CGNE) with eigenvector acceleration.
namespace	Chroma::LinOpMultiSysSolverEnv
	Registration aggregator.
namespace	Chroma::LinOpMultiSysSolverArrayEnv
	Registration aggregator.
namespace	Chroma::MdagMMultiSysSolverAccumulateEnv
	Registration aggregator.
namespace	Chroma::MdagMMultiSysSolverAccumulateArrayEnv
	Registration aggregator.
namespace	Chroma::MdagMMultiSysSolverEnv
	Registration aggregator.
namespace	Chroma::MdagMMultiSysSolverArrayEnv
	Registration aggregator.
namespace	Chroma::LinOpSysSolverEnv
	Registration aggregator.
namespace	Chroma::LinOpSysSolverArrayEnv
	Registration aggregator.
namespace	Chroma::MdagMSysSolverEnv
	Registration aggregator.
namespace	Chroma::MdagMSysSolverArrayEnv
	Registration aggregator.
namespace	Chroma::PolyPrecSysSolverEnv
	Registration aggregator.
Classes
class	Chroma::LinAlg::Vectors< T >
	Hold vectors. More...
class	Chroma::LinAlg::RitzPairs< T >
	Holds eigenvalues and eigenvectors. More...
class	Chroma::LinAlg::Matrix< T >
	This is a square matrix. More...
class	Chroma::LinAlg::VectorArrays< T >
	Hold vectors. More...
class	Chroma::LinAlg::RitzPairsArray< T >
	Holds eigenvalues and arrays of eigenvectors for use in 5D work. More...
struct	Chroma::MultiSysSolverCGParams
	Params for CG inverter. More...
class	Chroma::LinOpMultiSysSolverMR< T >
	Solve a MR system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMMultiSysSolverCG< T >
	Solve a CG2 system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMMultiSysSolverCGAccumulate< T >
	Solve a CG2 system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMMultiSysSolverCGAccumulateArray< T >
	Solve a CG system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMMultiSysSolverCGArray< T >
	Solve a CG system. Here, the operator is NOT assumed to be hermitian. More...
struct	Chroma::MultiSysSolverMRParams
	Params for MR inverter. More...
struct	Chroma::SysSolverBiCGStabParams
	Params for BiCGStab inverter. More...
struct	Chroma::SysSolverCGParams
	Params for CG inverter. More...
struct	Chroma::SysSolverEigCGParams
	Params for EigCG inverter. More...
class	Chroma::LinOpSysSolverBiCGStab< T >
	Solve a M*psi=chi linear system by BICGSTAB. More...
class	Chroma::LinOpSysSolverBiCRStab< T >
	Solve a M*psi=chi linear system by BICGSTAB. More...
class	Chroma::LinOpSysSolverCG< T >
	Solve a M*psi=chi linear system by CG2. More...
class	Chroma::LinOpSysSolverCGArray< T >
	Solve a M*psi=chi linear system by CG2. More...
class	Chroma::LinOpSysSolverCGTiming< T >
	Solve a M*psi=chi linear system by CG2. More...
class	Chroma::LinOpSysSolverEigCG< T >
	Solve a M*psi=chi linear system by EigCG with eigenvectors. More...
class	Chroma::LinOpSysSolverEigCGArray< T >
	Solve a M*psi=chi linear system by CG2 with eigenvectors. More...
class	Chroma::LinOpSysSolverIBiCGStab< T >
	Solve a M*psi=chi linear system by IBICGSTAB. More...
class	Chroma::LinOpSysSolverMR< T >
	Solve a M*psi=chi linear system by MR. More...
class	Chroma::LinOpSysSolverOptEigBiCG< T >
	Solve a M*psi=chi linear system by biCG with eigenvectors. More...
class	Chroma::LinOpSysSolverReliableBiCGStabClover
	Solve a system using Richardson iteration. More...
class	Chroma::LinOpSysSolverReliableCGClover
	Solve a system using Richardson iteration. More...
class	Chroma::LinOpSysSolverReliableIBiCGStabClover
	Solve a system using Richardson iteration. More...
class	Chroma::LinOpSysSolverRichardsonClover
	Solve a system using Richardson iteration. More...
class	Chroma::MdagMSysSolverBiCGStab< T >
	Solve a BiCGStab system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverCG< T >
	Solve a CG2 system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverCGArray< T >
	Solve a CG2 system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverCGLFClover
	Solve a system using CG iteration. More...
class	Chroma::MdagMSysSolverCGTimings< T >
	Solve a CG2 system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverQDPEigCG< T >
	Solve a M*psi=chi linear system by CG2 with eigenvectors. More...
class	Chroma::MdagMSysSolverIBiCGStab< T >
	Solve a IBiCGStab system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverMR< T >
	Solve a MR system. Here, the operator is NOT assumed to be hermitian. More...
class	Chroma::MdagMSysSolverOptEigCG< T >
	Solve a M*psi=chi linear system by CG2 with eigenvectors. More...
class	Chroma::MdagMSysSolverReliableBiCGStabClover
	Solve a system using Richardson iteration. More...
class	Chroma::MdagMSysSolverReliableCGClover
	Solve a system using Richardson iteration. More...
class	Chroma::MdagMSysSolverReliableIBiCGStabClover
	Solve a system using Richardson iteration. More...
class	Chroma::MdagMSysSolverRichardsonClover
	Solve a system using Richardson iteration. More...
struct	Chroma::SysSolverMRParams
	Params for MR inverter. More...
struct	Chroma::SysSolverOptEigBiCGParams
	Params for EigBiCG inverter. More...
struct	Chroma::SysSolverOptEigCGParams
	Params for EigCG inverter. More...
class	Chroma::PolyPrecSysSolverCG< T >
	Solve a PolyPrec*psi=chi linear system by CG1. More...
template<typename T>
SystemSolverResults_t	Chroma::InvBiCGStab (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdBiCGStab, int MaxBiCGStab, enum PlusMinus isign)
	Bi-CG stabilized.
template<typename T>
SystemSolverResults_t	Chroma::InvBiCRStab (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdBiCGStab, int MaxBiCGStab, enum PlusMinus isign)
	Bi-CG stabilized.
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperator< LatticeFermionF > &M, const LatticeFermionF &chi, LatticeFermionF &psi, const Real &RsdCG, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperator< LatticeFermionD > &M, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdCG, int MaxCG)
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperator< LatticeStaggeredFermionF > &M, const LatticeStaggeredFermionF &chi, LatticeStaggeredFermionF &psi, const Real &RsdCG, int MaxCG)
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperator< LatticeStaggeredFermionD > &M, const LatticeStaggeredFermionD &chi, LatticeStaggeredFermionD &psi, const Real &RsdCG, int MaxCG)
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperatorArray< LatticeFermionF > &M, const multi1d< LatticeFermionF > &chi, multi1d< LatticeFermionF > &psi, const Real &RsdCG, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
SystemSolverResults_t	Chroma::InvCG2 (const LinearOperatorArray< LatticeFermionD > &M, const multi1d< LatticeFermionD > &chi, multi1d< LatticeFermionD > &psi, const Real &RsdCG, int MaxCG)
template<typename T>
SystemSolverResults_t	Chroma::InvIBiCGStab (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdBiCGStab, int MaxBiCGStab, enum PlusMinus isign)
	Bi-CG stabilized.
template<typename T, typename C>
SystemSolverResults_t	Chroma::InvMR_a (const C &M, const T &chi, T &psi, const Real &MRovpar, const Real &RsdMR, int MaxMR, enum PlusMinus isign)
	Minimal-residual (MR) algorithm for a generic Linear Operator.
template<>
SystemSolverResults_t	Chroma::InvMR (const LinearOperator< LatticeFermion > &M, const LatticeFermion &chi, LatticeFermion &psi, const Real &MRovpar, const Real &RsdMR, int MaxMR, enum PlusMinus isign)
template<>
SystemSolverResults_t	Chroma::InvMR (const LinearOperator< LatticeStaggeredFermion > &M, const LatticeStaggeredFermion &chi, LatticeStaggeredFermion &psi, const Real &MRovpar, const Real &RsdMR, int MaxMR, enum PlusMinus isign)
template<typename T>
SystemSolverResults_t	Chroma::InvMR (const LinearOperator< T > &M, const T &chi, T &psi, const Real &MRovpar, const Real &RsdMR, int MaxMR, enum PlusMinus isign)
	Minimal-residual (MR) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::MInvMR_a (const LinearOperator< T > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdMR, int MaxMR, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvMR (const LinearOperator< LatticeFermion > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdMR, int MaxMR, int &n_count)
template<>
void	Chroma::MInvMR (const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdMR, int MaxMR, int &n_count)
template<typename T>
void	Chroma::MInvMR (const LinearOperator< T > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<typename T, typename P, typename Q>
void	Chroma::MInvMR (const DiffLinearOperator< T, P, Q > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
void	Chroma::normGramSchmidt (multi1d< LatticeFermionF > &vec, int f, int t, const Subset &sub)
	Gram-Schmidt with normalization.
void	Chroma::normGramSchmidt (multi1d< LatticeFermionD > &vec, int f, int t, const Subset &sub)
void	Chroma::normGramSchmidt (multi1d< LatticeStaggeredFermionF > &vec, int f, int t, const Subset &sub)
void	Chroma::normGramSchmidt (multi1d< LatticeStaggeredFermionD > &vec, int f, int t, const Subset &sub)
void	Chroma::normGramSchmidt (multi2d< LatticeFermionF > &vec, int f, int t, const Subset &sub)
void	Chroma::normGramSchmidt (multi2d< LatticeFermionD > &vec, int f, int t, const Subset &sub)
SystemSolverResults_t	Chroma::InvBiCGStabReliable (const LinearOperator< LatticeFermionF > &A, const LatticeFermionF &chi, LatticeFermionF &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
	Bi-CG stabilized.
SystemSolverResults_t	Chroma::InvBiCGStabReliable (const LinearOperator< LatticeFermionD > &A, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
SystemSolverResults_t	Chroma::InvBiCGStabReliable (const LinearOperator< LatticeFermionD > &A, const LinearOperator< LatticeFermionF > &AF, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
SystemSolverResults_t	Chroma::InvCGReliable (const LinearOperator< LatticeFermionF > &A, const LatticeFermionF &chi, LatticeFermionF &psi, const Real &RsdCG, const Real &Delta, int MaxCG)
	Bi-CG stabilized.
SystemSolverResults_t	Chroma::InvCGReliable (const LinearOperator< LatticeFermionD > &A, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdCG, const Real &Delta, int MaxCG)
SystemSolverResults_t	Chroma::InvCGReliable (const LinearOperator< LatticeFermionD > &A, const LinearOperator< LatticeFermionF > &AF, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdCG, const Real &Delta, int MaxCG)
SystemSolverResults_t	Chroma::InvIBiCGStabReliable (const LinearOperator< LatticeFermionF > &A, const LatticeFermionF &chi, LatticeFermionF &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
	Bi-CG stabilized.
SystemSolverResults_t	Chroma::InvIBiCGStabReliable (const LinearOperator< LatticeFermionD > &A, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
SystemSolverResults_t	Chroma::InvIBiCGStabReliable (const LinearOperator< LatticeFermionD > &A, const LinearOperator< LatticeFermionF > &AF, const LatticeFermionD &chi, LatticeFermionD &psi, const Real &RsdBiCGStab, const Real &Delta, int MaxBiCGStab, enum PlusMinus isign)
Typedefs
typedef SingletonHolder< ObjectFactory< LinOpMultiSystemSolver< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), LinOpMultiSystemSolver< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::TheLinOpFermMultiSystemSolverFactory
	LinOp system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< LinOpMultiSystemSolver< LatticeStaggeredFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), LinOpMultiSystemSolver< LatticeStaggeredFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), StringFactoryError > >	Chroma::TheLinOpStagFermMultiSystemSolverFactory
	LinOp system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolverAccumulate< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), MdagMMultiSystemSolverAccumulate< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermMultiSystemSolverAccumulateFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolverAccumulateArray< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), MdagMMultiSystemSolverAccumulateArray< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermMultiSystemSolverAccumulateArrayFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolverAccumulate< LatticeStaggeredFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), MdagMMultiSystemSolverAccumulate< LatticeStaggeredFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), StringFactoryError > >	Chroma::TheMdagMStagFermMultiSystemSolverAccumulateFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolver< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), MdagMMultiSystemSolver< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermMultiSystemSolverFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolverArray< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), MdagMMultiSystemSolverArray< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermMultiSystemSolverArrayFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMMultiSystemSolver< LatticeStaggeredFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), MdagMMultiSystemSolver< LatticeStaggeredFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), StringFactoryError > >	Chroma::TheMdagMStagFermMultiSystemSolverFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< LinOpSystemSolver< LatticeFermion >, std::string, TYPELIST_4(XMLReader &, const std::string &, FSHandle, Handle< LinearOperator< LatticeFermion > >), LinOpSystemSolver< LatticeFermion > ()(XMLReader &, const std::string &, FSHandle, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::TheLinOpFermSystemSolverFactory
	LinOp system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< LinOpSystemSolverArray< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), LinOpSystemSolverArray< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), StringFactoryError > >	Chroma::TheLinOpFermSystemSolverArrayFactory
	LinOp system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< LinOpSystemSolver< LatticeStaggeredFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), LinOpSystemSolver< LatticeStaggeredFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), StringFactoryError > >	Chroma::TheLinOpStagFermSystemSolverFactory
	LinOp system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMSystemSolver< LatticeFermion >, std::string, TYPELIST_4(XMLReader &, const std::string &, FactoryEnv::FSHandle, Handle< LinearOperator< LatticeFermion > >), MdagMSystemSolver< LatticeFermion > ()(XMLReader &, const std::string &, FactoryEnv::FSHandle, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermSystemSolverFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMSystemSolverArray< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), MdagMSystemSolverArray< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperatorArray< LatticeFermion > >), StringFactoryError > >	Chroma::TheMdagMFermSystemSolverArrayFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< MdagMSystemSolver< LatticeStaggeredFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), MdagMSystemSolver< LatticeStaggeredFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeStaggeredFermion > >), StringFactoryError > >	Chroma::TheMdagMStagFermSystemSolverFactory
	MdagM system solver factory (foundry).
typedef SingletonHolder< ObjectFactory< PolyPrecSystemSolver< LatticeFermion >, std::string, TYPELIST_3(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), PolyPrecSystemSolver< LatticeFermion > ()(XMLReader &, const std::string &, Handle< LinearOperator< LatticeFermion > >), StringFactoryError > >	Chroma::ThePolyPrecFermSystemSolverFactory
	PolyPrec system solver factory (foundry).
Functions
template<typename T>
void	Chroma::InvRelCG1_a (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::InvRelCG1 (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::InvRelCG2_a (const LinearOperator< T > &M, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::InvRelCG2 (const LinearOperator< T > &M, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Relaxed Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
SystemSolverResults_t	Chroma::InvCG1_a (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int MinCG=0)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
SystemSolverResults_t	Chroma::InvCG1 (const LinearOperator< T > &A, const T &chi, T &psi, const Real &RsdCG, int MaxCG, int MinCG=0)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::InvCG1_a (const LinearOperatorArray< T > &A, const multi1d< T > &chi, multi1d< T > &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::InvCG1 (const LinearOperatorArray< T > &A, const multi1d< T > &chi, multi1d< T > &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T, typename RT>
SystemSolverResults_t	Chroma::InvCG2_a (const LinearOperator< T > &M, const T &chi, T &psi, const RT &RsdCG, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T, typename C>
SystemSolverResults_t	Chroma::InvCG2_a (const C &M, const multi1d< T > &chi, multi1d< T > &psi, const Real &RsdCG, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
SystemSolverResults_t	Chroma::InvCG2_timings_a (const LinearOperator< T > &M, const T &chi, T &psi, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
SystemSolverResults_t	Chroma::InvCG2_timings (const LinearOperator< T > &M, const T &chi, T &psi, int MaxCG)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
template<typename T>
void	Chroma::MInvRelCG_a (const LinearOperator< T > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<typename T>
void	Chroma::MInvCG_a (const LinearOperator< T > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvCG (const LinearOperator< LatticeFermion > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<>
void	Chroma::MInvCG (const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG (const LinearOperator< T > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T, typename P, typename Q>
void	Chroma::MInvCG (const DiffLinearOperator< T, P, Q > &A, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG2_a (const LinearOperator< T > &M, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvCG2 (const LinearOperator< LatticeFermion > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<>
void	Chroma::MInvCG2 (const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const LatticeFermion &chi, multi1d< LatticeFermion > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG2 (const LinearOperator< T > &M, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T, typename P, typename Q>
void	Chroma::MInvCG2 (const DiffLinearOperator< T, P, Q > &M, const T &chi, multi1d< T > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG2Accum_a (const LinearOperator< T > &M, const T &chi, T &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvCG2Accum (const LinearOperator< LatticeFermion > &M, const LatticeFermion &chi, LatticeFermion &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, int MaxCG, int &n_count)
template<>
void	Chroma::MInvCG2Accum (const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const LatticeFermion &chi, LatticeFermion &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG2Accum (const LinearOperator< T > &M, const T &chi, T &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, int MaxCG, int &n_count)
template<typename T, typename P, typename Q>
void	Chroma::MInvCG2Accum (const DiffLinearOperator< T, P, Q > &M, const T &chi, T &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, int MaxCG, int &n_count)
template<typename T, typename C>
void	Chroma::MInvCGAccum_a (const C &A, const multi1d< T > &chi, multi1d< T > &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, const int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvCGAccum (const LinearOperatorArray< LatticeFermion > &M, const multi1d< LatticeFermion > &chi, multi1d< LatticeFermion > &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, const int MaxCG, int &n_count)
template<>
void	Chroma::MInvCGAccum (const DiffLinearOperatorArray< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const multi1d< LatticeFermion > &chi, multi1d< LatticeFermion > &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, const int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCGAccum (const LinearOperatorArray< T > &A, const multi1d< T > &chi, multi1d< T > &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, const int MaxCG, int &n_count)
template<typename T, typename P, typename Q>
void	Chroma::MInvCGAccum (const DiffLinearOperatorArray< T, P, Q > &A, const multi1d< T > &chi, multi1d< T > &psi, const Real &norm, const multi1d< Real > &residues, const multi1d< Real > &poles, const Real &RsdCG, const int MaxCG, int &n_template)
template<typename T, typename C>
void	Chroma::MInvCG_a (const C &A, const multi1d< T > &chi, multi1d< multi1d< T > > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
	Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.
template<>
void	Chroma::MInvCG (const LinearOperatorArray< LatticeFermion > &M, const multi1d< LatticeFermion > &chi, multi1d< multi1d< LatticeFermion > > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<>
void	Chroma::MInvCG (const DiffLinearOperatorArray< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &M, const multi1d< LatticeFermion > &chi, multi1d< multi1d< LatticeFermion > > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T>
void	Chroma::MInvCG (const LinearOperatorArray< T > &A, const multi1d< T > &chi, multi1d< multi1d< T > > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
template<typename T, typename P, typename Q>
void	Chroma::MInvCG (const DiffLinearOperatorArray< T, P, Q > &A, const multi1d< T > &chi, multi1d< multi1d< T > > &psi, const multi1d< Real > &shifts, const multi1d< Real > &RsdCG, int MaxCG, int &n_count)
void	Chroma::read (XMLReader &xml, const string &path, MultiSysSolverCGParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const MultiSysSolverCGParams &param)
void	Chroma::read (XMLReader &xml, const string &path, MultiSysSolverMRParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const MultiSysSolverMRParams &param)
template<typename T>
void	Chroma::normGramSchmidt_T (multi1d< T > &vec, int f, int t, const Subset &sub)
	Gram-Schmidt with normalization.
template<typename T>
void	Chroma::normGramSchmidtArray_T (multi2d< T > &vec, int f, int t, const Subset &sub)
	Gram-Schmidt with normalization.
void	Chroma::read (XMLReader &xml, const string &path, SysSolverBiCGStabParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverBiCGStabParams &param)
void	Chroma::read (XMLReader &xml, const string &path, SysSolverCGParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverCGParams &param)
void	Chroma::read (XMLReader &xml, const string &path, SysSolverEigCGParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverEigCGParams &param)
void	Chroma::read (XMLReader &xml, const string &path, SysSolverMRParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverMRParams &param)
void	Chroma::read (XMLReader &xml, const string &path, SysSolverOptEigBiCGParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverOptEigBiCGParams &param)
void	Chroma::read (XMLReader &xml, const string &path, SysSolverOptEigCGParams &param)
void	Chroma::write (XMLWriter &xml, const string &path, const SysSolverOptEigCGParams &param)
void	InvCG2EvenOddPrecWilsLinOp (const WilsonDslash &D, const LFerm &chi, LFerm &psi, const LScal &mass, const LScal &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
void	InvCG2EvenOddPrecWilsLinOpTHack (const WilsonDslash &D, const LFerm &chi, LFerm &psi, const LScal &mass, const LScal &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.
void	WlInvCG2 (const LinearOperator &M, const LatticeFermion &chi, LatticeFermion &psi, const Real &RsdCG, int MaxCG, int &n_count)
	Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

Detailed Description

Various inverters for fermion linear operators

Typedef Documentation

typedef SingletonHolder< ObjectFactory<LinOpMultiSystemSolver<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), LinOpMultiSystemSolver<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::TheLinOpFermMultiSystemSolverFactory

LinOp system solver factory (foundry).

Definition at line 28 of file multi_syssolver_linop_factory.h.

typedef SingletonHolder< ObjectFactory<LinOpSystemSolverArray<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), LinOpSystemSolverArray<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), StringFactoryError> > Chroma::TheLinOpFermSystemSolverArrayFactory

LinOp system solver factory (foundry).

Definition at line 77 of file syssolver_linop_factory.h.

typedef SingletonHolder< ObjectFactory<LinOpSystemSolver<LatticeFermion>, std::string, TYPELIST_4(XMLReader&, const std::string&, FSHandle, Handle< LinearOperator<LatticeFermion> >), LinOpSystemSolver<LatticeFermion>* (*)(XMLReader&, const std::string&, FSHandle, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::TheLinOpFermSystemSolverFactory

LinOp system solver factory (foundry).

Definition at line 40 of file syssolver_linop_factory.h.

typedef SingletonHolder< ObjectFactory<LinOpMultiSystemSolver<LatticeStaggeredFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), LinOpMultiSystemSolver<LatticeStaggeredFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), StringFactoryError> > Chroma::TheLinOpStagFermMultiSystemSolverFactory

LinOp system solver factory (foundry).

Definition at line 56 of file multi_syssolver_linop_factory.h.

typedef SingletonHolder< ObjectFactory<LinOpSystemSolver<LatticeStaggeredFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), LinOpSystemSolver<LatticeStaggeredFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), StringFactoryError> > Chroma::TheLinOpStagFermSystemSolverFactory

LinOp system solver factory (foundry).

Definition at line 91 of file syssolver_linop_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolverAccumulateArray<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), MdagMMultiSystemSolverAccumulateArray<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermMultiSystemSolverAccumulateArrayFactory

MdagM system solver factory (foundry).

Definition at line 41 of file multi_syssolver_mdagm_accumulate_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolverAccumulate<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), MdagMMultiSystemSolverAccumulate<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermMultiSystemSolverAccumulateFactory

MdagM system solver factory (foundry).

Definition at line 28 of file multi_syssolver_mdagm_accumulate_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolverArray<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), MdagMMultiSystemSolverArray<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermMultiSystemSolverArrayFactory

MdagM system solver factory (foundry).

Definition at line 41 of file multi_syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolver<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), MdagMMultiSystemSolver<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermMultiSystemSolverFactory

MdagM system solver factory (foundry).

Definition at line 28 of file multi_syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMSystemSolverArray<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), MdagMSystemSolverArray<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperatorArray<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermSystemSolverArrayFactory

MdagM system solver factory (foundry).

Definition at line 75 of file syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMSystemSolver<LatticeFermion>, std::string, TYPELIST_4(XMLReader&, const std::string&, FactoryEnv::FSHandle, Handle< LinearOperator<LatticeFermion> >), MdagMSystemSolver<LatticeFermion>* (*)(XMLReader&, const std::string&, FactoryEnv::FSHandle, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::TheMdagMFermSystemSolverFactory

MdagM system solver factory (foundry).

Definition at line 40 of file syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolverAccumulate<LatticeStaggeredFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), MdagMMultiSystemSolverAccumulate<LatticeStaggeredFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), StringFactoryError> > Chroma::TheMdagMStagFermMultiSystemSolverAccumulateFactory

MdagM system solver factory (foundry).

Definition at line 54 of file multi_syssolver_mdagm_accumulate_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMMultiSystemSolver<LatticeStaggeredFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), MdagMMultiSystemSolver<LatticeStaggeredFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), StringFactoryError> > Chroma::TheMdagMStagFermMultiSystemSolverFactory

MdagM system solver factory (foundry).

Definition at line 54 of file multi_syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<MdagMSystemSolver<LatticeStaggeredFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), MdagMSystemSolver<LatticeStaggeredFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeStaggeredFermion> >), StringFactoryError> > Chroma::TheMdagMStagFermSystemSolverFactory

MdagM system solver factory (foundry).

Definition at line 88 of file syssolver_mdagm_factory.h.

typedef SingletonHolder< ObjectFactory<PolyPrecSystemSolver<LatticeFermion>, std::string, TYPELIST_3(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), PolyPrecSystemSolver<LatticeFermion>* (*)(XMLReader&, const std::string&, Handle< LinearOperator<LatticeFermion> >), StringFactoryError> > Chroma::ThePolyPrecFermSystemSolverFactory

PolyPrec system solver factory (foundry).

Definition at line 27 of file syssolver_polyprec_factory.h.

Function Documentation

template<typename T>

SystemSolverResults_t Chroma::InvBiCGStab	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdBiCGStab,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Bi-CG stabilized.

SystemSolverResults_t Chroma::InvBiCGStabReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LinearOperator< LatticeFermionF > &	AF,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Definition at line 324 of file reliable_bicgstab.cc.

SystemSolverResults_t Chroma::InvBiCGStabReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Definition at line 310 of file reliable_bicgstab.cc.

SystemSolverResults_t Chroma::InvBiCGStabReliable	(	const LinearOperator< LatticeFermionF > &	A,
		const LatticeFermionF &	chi,
		LatticeFermionF &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Bi-CG stabilized.

Definition at line 296 of file reliable_bicgstab.cc.

Referenced by Chroma::MdagMSysSolverReliableBiCGStabClover::operator()(), and Chroma::LinOpSysSolverReliableBiCGStabClover::operator()().

template<typename T>

SystemSolverResults_t Chroma::InvBiCRStab	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdBiCGStab,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Bi-CG stabilized.

template<typename T>

void Chroma::InvCG1	(	const LinearOperatorArray< T > &	A,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A is hermitian

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - A . Psi[0] ; Initial residual p[1] := r[0] ; Initial direction IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k-1]|**2 / <A p[k],p[k]> ; Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] A. p[k] ; New residual IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged? b[k+1] := |r[k]|**2 / |r[k-1]|**2 ; p[k+1] := r[k] + b[k+1] p[k]; New direction

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: +

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

template<typename T>

SystemSolverResults_t Chroma::InvCG1	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int	MinCG = `0`
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A is hermitian

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - A. Psi[0] ; Initial residual p[1] := r[0] ; Initial direction IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k-1]|**2 / <p[k],A p[k]> ; Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] A p[k] ; New residual IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged? b[k+1] := |r[k]|**2 / |r[k-1]|**2 ; p[k+1] := r[k] + b[k+1] p[k]; New direction

Arguments:

Parameters:

	A	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)

Returns:: res System solver results

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > ap Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

template<typename T>

void Chroma::InvCG1_a	(	const LinearOperatorArray< T > &	A,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A is hermitian

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - M^dag . M . Psi[0] ; Initial residual p[1] := r[0] ; Initial direction IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k-1]|**2 / <Mp[k],Mp[k]> ; Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] M^dag . M . p[k] ; New residual IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged? b[k+1] := |r[k]|**2 / |r[k-1]|**2 ; p[k+1] := r[k] + b[k+1] p[k]; New direction

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 68 of file invcg1_array.cc.

References END_CODE, Chroma::PLUS, START_CODE, and Chroma::LinearOperatorArray< T >::subset().

template<typename T>

SystemSolverResults_t Chroma::InvCG1_a	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int	MinCG = `0`
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A is Hermitian Positive Definite

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - A. Psi[0] ; Initial residual p[1] := r[0] ; Initial direction IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k-1]|**2 / <p[k],Ap[k]> ; Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] A. p[k] ; New residual IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged? b[k+1] := |r[k]|**2 / |r[k-1]|**2 ; p[k+1] := r[k] + b[k+1] p[k]; New direction

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Rea/Write)
	MaxCG	Maximum CG iterations (Read)

Returns:: res System solver results

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > ap Temporary for A.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 70 of file invcg1.cc.

References END_CODE, Chroma::SystemSolverResults_t::n_count, Chroma::PLUS, Chroma::SystemSolverResults_t::resid, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::InvCG1().

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperatorArray< LatticeFermionD > &	M,
		const multi1d< LatticeFermionD > &	chi,
		multi1d< LatticeFermionD > &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Definition at line 293 of file invcg2_array.cc.

References Chroma::InvCG2_a().

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperator< LatticeStaggeredFermionD > &	M,
		const LatticeStaggeredFermionD &	chi,
		LatticeStaggeredFermionD &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Definition at line 287 of file invcg2.cc.

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperator< LatticeStaggeredFermionF > &	M,
		const LatticeStaggeredFermionF &	chi,
		LatticeStaggeredFermionF &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Definition at line 276 of file invcg2.cc.

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperator< LatticeFermionD > &	M,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Definition at line 259 of file invcg2.cc.

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperatorArray< LatticeFermionF > &	M,
		const multi1d< LatticeFermionF > &	chi,
		multi1d< LatticeFermionF > &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 282 of file invcg2_array.cc.

References Chroma::InvCG2_a().

SystemSolverResults_t Chroma::InvCG2	(	const LinearOperator< LatticeFermionF > &	M,
		const LatticeFermionF &	chi,
		LatticeFermionF &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 241 of file invcg2.cc.

template<typename T, typename C>

SystemSolverResults_t Chroma::InvCG2_a	(	const C &	M,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	RsdCG,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)

Returns:: res System solver results

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 69 of file invcg2_array.cc.

References END_CODE, Chroma::MINUS, Chroma::SystemSolverResults_t::n_count, Chroma::PLUS, Chroma::SystemSolverResults_t::resid, and START_CODE.

template<typename T, typename RT>

SystemSolverResults_t Chroma::InvCG2_a	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const RT &	RsdCG,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)

Returns:: res System solver results

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 73 of file invcg2.cc.

References END_CODE, Chroma::MINUS, Chroma::SystemSolverResults_t::n_count, Chroma::LinearOperator< T >::nFlops(), Chroma::PLUS, Chroma::SystemSolverResults_t::resid, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::InvCG2().

template<typename T>

SystemSolverResults_t Chroma::InvCG2_timings	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

template<typename T>

SystemSolverResults_t Chroma::InvCG2_timings_a	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		int	MaxCG
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 71 of file invcg2_timing_hacks.cc.

Referenced by Chroma::InvCG2_timings().

void InvCG2EvenOddPrecWilsLinOp	(	const WilsonDslash &	D,
		const LFerm &	chi,
		LFerm &	psi,
		const LScal &	mass,
		const LScal &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 67 of file invcg2_prec_wilson.cc.

References AT_REAL, Chroma::DOUBLE, FIRST_ELEM, Chroma::MINUS, Chroma::PLUS, and vaxpy3_norm().

void InvCG2EvenOddPrecWilsLinOpTHack	(	const WilsonDslash &	D,
		const LFerm &	chi,
		LFerm &	psi,
		const LScal &	mass,
		const LScal &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 67 of file invcg2_timing_hacks_2.cc.

References AT_REAL, Chroma::DOUBLE, FIRST_ELEM, Chroma::MINUS, Chroma::PLUS, and vaxpy3_norm().

SystemSolverResults_t Chroma::InvCGReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LinearOperator< LatticeFermionF > &	AF,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdCG,
		const Real &	Delta,
		int	MaxCG
	)

Definition at line 219 of file reliable_cg.cc.

SystemSolverResults_t Chroma::InvCGReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdCG,
		const Real &	Delta,
		int	MaxCG
	)

Definition at line 207 of file reliable_cg.cc.

SystemSolverResults_t Chroma::InvCGReliable	(	const LinearOperator< LatticeFermionF > &	A,
		const LatticeFermionF &	chi,
		LatticeFermionF &	psi,
		const Real &	RsdCG,
		const Real &	Delta,
		int	MaxCG
	)

Bi-CG stabilized.

Definition at line 194 of file reliable_cg.cc.

Referenced by Chroma::MdagMSysSolverReliableCGClover::operator()(), and Chroma::LinOpSysSolverReliableCGClover::operator()().

template<typename T>

SystemSolverResults_t Chroma::InvIBiCGStab	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdBiCGStab,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Bi-CG stabilized.

SystemSolverResults_t Chroma::InvIBiCGStabReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LinearOperator< LatticeFermionF > &	AF,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Definition at line 467 of file reliable_ibicgstab.cc.

SystemSolverResults_t Chroma::InvIBiCGStabReliable	(	const LinearOperator< LatticeFermionD > &	A,
		const LatticeFermionD &	chi,
		LatticeFermionD &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Definition at line 453 of file reliable_ibicgstab.cc.

SystemSolverResults_t Chroma::InvIBiCGStabReliable	(	const LinearOperator< LatticeFermionF > &	A,
		const LatticeFermionF &	chi,
		LatticeFermionF &	psi,
		const Real &	RsdBiCGStab,
		const Real &	Delta,
		int	MaxBiCGStab,
		enum PlusMinus	isign
	)

Bi-CG stabilized.

Definition at line 439 of file reliable_ibicgstab.cc.

Referenced by Chroma::MdagMSysSolverReliableIBiCGStabClover::operator()(), and Chroma::LinOpSysSolverReliableIBiCGStabClover::operator()().

template<>

SystemSolverResults_t Chroma::InvMR	(	const LinearOperator< LatticeStaggeredFermion > &	M,
		const LatticeStaggeredFermion &	chi,
		LatticeStaggeredFermion &	psi,
		const Real &	MRovpar,
		const Real &	RsdMR,
		int	MaxMR,
		enum PlusMinus	isign
	)

Definition at line 201 of file invmr.cc.

References Chroma::InvMR_a().

template<>

SystemSolverResults_t Chroma::InvMR	(	const LinearOperator< LatticeFermion > &	M,
		const LatticeFermion &	chi,
		LatticeFermion &	psi,
		const Real &	MRovpar,
		const Real &	RsdMR,
		int	MaxMR,
		enum PlusMinus	isign
	)

Definition at line 186 of file invmr.cc.

References Chroma::InvMR_a().

Referenced by Chroma::MdagMSysSolverMR< T >::operator()(), Chroma::LinOpSysSolverMR< T >::operator()(), and Chroma::Ovlap4DQprop::operator()().

template<typename T>

SystemSolverResults_t Chroma::InvMR	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const Real &	MRovpar,
		const Real &	RsdMR,
		int	MaxMR,
		enum PlusMinus	isign
	)

Minimal-residual (MR) algorithm for a generic Linear Operator.

This subroutine uses the Minimal Residual (MR) algorithm to determine the solution of the set of linear equations. Here we allow M to be nonhermitian.

Chi = M . Psi

Algorithm:

Psi[0] Argument r[0] := Chi - M . Psi[0] ; Initial residual IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO MR iterations a[k-1] := <M.r[k-1],r[k-1]> / <M.r[k-1],M.r[k-1]> ; ap[k-1] := MRovpar * a[k] ; Overrelaxtion step Psi[k] += ap[k-1] r[k-1] ; New solution vector r[k] -= ap[k-1] A . r[k-1] ; New residual IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged?

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	MR residual accuracy (Read)
	MRovpar	Overrelaxation parameter (Read)
	MaxCG	Maximum MR iterations (Read)

Local Variables:

r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k MR iteration counter a a[k] d < M.r[k], M.r[k] > R_Aux Temporary for M.Psi Mr Temporary for M.r

Global Variables:

MaxCG Maximum number of MR iterations allowed RsdCG Maximum acceptable MR residual (relative to source)

Subroutines:

M Apply matrix to vector

template<typename T, typename C>

SystemSolverResults_t Chroma::InvMR_a	(	const C &	M,
		const T &	chi,
		T &	psi,
		const Real &	MRovpar,
		const Real &	RsdMR,
		int	MaxMR,
		enum PlusMinus	isign
	)

Minimal-residual (MR) algorithm for a generic Linear Operator.

This subroutine uses the Minimal Residual (MR) algorithm to determine the solution of the set of linear equations. Here we allow M to be nonhermitian.

Chi = M . Psi

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	MR residual accuracy (Read)
	MRovpar	Overrelaxation parameter (Read)
	MaxMR	Maximum MR iterations (Read)

Local Variables:

r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k MR iteration counter a a[k] d < M.r[k], M.r[k] > R_Aux Temporary for M.Psi Mr Temporary for M.r

Global Variables:

MaxMR Maximum number of MR iterations allowed RsdCG Maximum acceptable MR residual (relative to source)

Subroutines:

M Apply matrix to vector

Definition at line 67 of file invmr.cc.

References END_CODE, Chroma::SystemSolverResults_t::n_count, Chroma::SystemSolverResults_t::resid, START_CODE, and T.

Referenced by Chroma::InvMR().

template<typename T>

void Chroma::InvRelCG1	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A is hermitian

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - A. Psi[0] ; Initial residual p[1] := r[0] ; Initial direction c := cp := || r[0] ||^2 zeta := 1/c;

IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations inner_tol := RsdCG*||chi||*||p||*sqrt(zeta)

q := A(inner_tol) p

a[k] := |r[k-1]|**2 / <q,p[k]> ; Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] q ; New residual c := <r , r> zeta := zeta + 1/c

b[k+1] := |r[k]|**2 / |r[k-1]|**2 = c/cp; p[k+1] := r[k] + b[k+1] p[k]; New direction

cp := c IF ( c < RsdCG^2 || chi || ) RETURN

Arguments:

Parameters:

	A	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > ap Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

template<typename T>

void Chroma::InvRelCG1_a	(	const LinearOperator< T > &	A,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

Definition at line 21 of file inv_rel_cg1.cc.

References END_CODE, Chroma::PLUS, q, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::InvRelCG1().

template<typename T>

void Chroma::InvRelCG2	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Relaxed Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = M^{dag}M . Psi

Algorithm:

Psi[0] := initial guess; Linear interpolation (argument) r[0] := Chi - M^dag . M . Psi[0] ; Initial residual p[1] := r[0] ; Initial direction c = cp := || r[0] ||^2 zeta := 1/c;

IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged?

FOR k FROM 1 TO MaxCG DO CG iterations

inner_tol := RsdCG*||chi||*||p||*sqrt(zeta)/2; q := M^{dag}(tol) M(tol) p; a[k] := c / <q , p> Psi[k] += a[k] p[k] ; New solution vector r[k] -= a[k] q; New residual c := || r[k]^2 || zeta = zeta + 1/c; b[k+1] := |r[k]|**2 / |r[k-1]|**2 = c/cp; p[k+1] := r[k] + b[k+1] p[k]; New direction cp := c; IF |r[k]| <= RsdCG |Chi| THEN RETURN; Converged?

Arguments:

Parameters:

	M	ApproxLinear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

template<typename T>

void Chroma::InvRelCG2_a	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 65 of file inv_rel_cg2.cc.

References Chroma::MINUS, Chroma::PLUS, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::InvRelCG2().

template<typename T, typename P, typename Q>

void Chroma::MInvCG	(	const DiffLinearOperatorArray< T, P, Q > &	A,
		const multi1d< T > &	chi,
		multi1d< multi1d< T > > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<typename T>

void Chroma::MInvCG	(	const LinearOperatorArray< T > &	A,
		const multi1d< T > &	chi,
		multi1d< multi1d< T > > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<>

void Chroma::MInvCG	(	const DiffLinearOperatorArray< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const multi1d< LatticeFermion > &	chi,
		multi1d< multi1d< LatticeFermion > > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 454 of file minvcg_array.cc.

References Chroma::MInvCG_a(), and Chroma::psi().

template<>

void Chroma::MInvCG	(	const LinearOperatorArray< LatticeFermion > &	M,
		const multi1d< LatticeFermion > &	chi,
		multi1d< multi1d< LatticeFermion > > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 440 of file minvcg_array.cc.

References Chroma::MInvCG_a(), and Chroma::psi().

template<typename T, typename P, typename Q>

void Chroma::MInvCG	(	const DiffLinearOperator< T, P, Q > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<typename T>

void Chroma::MInvCG	(	const LinearOperator< T > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<>

void Chroma::MInvCG	(	const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 422 of file minvcg.cc.

References Chroma::MInvCG_a().

template<>

void Chroma::MInvCG	(	const LinearOperator< LatticeFermion > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 408 of file minvcg.cc.

References Chroma::MInvCG_a().

Referenced by Chroma::OverlapFermActBase::multiQprop(), and Chroma::MdagMMultiSysSolverCGArray< T >::operator()().

template<typename T, typename P, typename Q>

void Chroma::MInvCG2	(	const DiffLinearOperator< T, P, Q > &	M,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<typename T>

void Chroma::MInvCG2	(	const LinearOperator< T > &	M,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<>

void Chroma::MInvCG2	(	const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 388 of file minvcg2.cc.

References Chroma::MInvCG2_a().

template<>

void Chroma::MInvCG2	(	const LinearOperator< LatticeFermion > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 368 of file minvcg2.cc.

References Chroma::MInvCG2_a().

Referenced by Chroma::MdagMMultiSysSolverCG< T >::operator()().

template<typename T>

void Chroma::MInvCG2_a	(	const LinearOperator< T > &	M,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = M^ M . Psi

Algorithm:

Psi[0] := 0; Zeroed r[0] := Chi; Initial residual p[1] := Chi ; Initial direction b[0] := |r[0]|**2 / <M p[0], M p[0]> ; z[0] := 1 / (1 - (shift - shift(0))*b) bs[0] := b[0] * z[0] r[1] += b[k] M^ M . p[0] ; New residual Psi[1] = - b[k] p[k] ; Starting solution vector IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k]|**2 / |r[k-1]|**2 ; p[k] := r[k] + a[k] p[k-1]; New direction b[k+1] := |r[k]|**2 / <M p[k], Mp[k]> ; r[k+1] += b[k+1] M^ M . p[k] ; New residual Psi[k+1] -= b[k+1] p[k] ; New solution vector IF |[k+1]| <= RsdCG |Chi| THEN RETURN; Converged?

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 76 of file minvcg2.cc.

References Chroma::chi_norm, END_CODE, Chroma::MINUS, Chroma::LinearOperator< T >::nFlops(), Chroma::PLUS, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::MInvCG2().

template<typename T, typename P, typename Q>

void Chroma::MInvCG2Accum	(	const DiffLinearOperator< T, P, Q > &	M,
		const T &	chi,
		T &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<typename T>

void Chroma::MInvCG2Accum	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<>

void Chroma::MInvCG2Accum	(	const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const LatticeFermion &	chi,
		LatticeFermion &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 389 of file minvcg2_accum.cc.

References Chroma::MInvCG2Accum_a().

template<>

void Chroma::MInvCG2Accum	(	const LinearOperator< LatticeFermion > &	M,
		const LatticeFermion &	chi,
		LatticeFermion &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Definition at line 373 of file minvcg2_accum.cc.

References Chroma::MInvCG2Accum_a().

Referenced by Chroma::MdagMMultiSysSolverCGAccumulate< T >::operator()().

template<typename T>

void Chroma::MInvCG2Accum_a	(	const LinearOperator< T > &	M,
		const T &	chi,
		T &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = M^ M . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 73 of file minvcg2_accum.cc.

References Chroma::chi_norm, END_CODE, Chroma::MINUS, Chroma::LinearOperator< T >::nFlops(), Chroma::PLUS, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::MInvCG2Accum().

template<typename T, typename C>

void Chroma::MInvCG_a	(	const C &	A,
		const multi1d< T > &	chi,
		multi1d< multi1d< T > > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Psi[0] := 0; Zeroed r[0] := Chi; Initial residual p[1] := Chi ; Initial direction b[0] := |r[0]|**2 / <p[0],Ap[0]> ; z[0] := 1 / (1 - (shift - shift(0))*b) bs[0] := b[0] * z[0] r[1] += b[k] A . p[0] ; New residual Psi[1] = - b[k] p[k] ; Starting solution vector IF |r[0]| <= RsdCG |Chi| THEN RETURN; Converged? FOR k FROM 1 TO MaxCG DO CG iterations a[k] := |r[k]|**2 / |r[k-1]|**2 ; p[k] := r[k] + a[k] p[k-1]; New direction b[k+1] := |r[k]|**2 / <p[k],Ap[k]> ; r[k+1] += b[k+1] A . p[k] ; New residual Psi[k+1] -= b[k+1] p[k] ; New solution vector IF |[k+1]| <= RsdCG |Chi| THEN RETURN; Converged?

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 75 of file minvcg_array.cc.

References Chroma::chi_norm, END_CODE, Chroma::PLUS, Chroma::psi(), and START_CODE.

template<typename T>

void Chroma::MInvCG_a	(	const LinearOperator< T > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 73 of file minvcg.cc.

References Chroma::chi_norm, END_CODE, Chroma::LinearOperator< T >::nFlops(), Chroma::PLUS, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::MInvCG().

template<typename T, typename P, typename Q>

void Chroma::MInvCGAccum	(	const DiffLinearOperatorArray< T, P, Q > &	A,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		const int	MaxCG,
		int &	n_template
	)

template<typename T>

void Chroma::MInvCGAccum	(	const LinearOperatorArray< T > &	A,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		const int	MaxCG,
		int &	n_count
	)

template<>

void Chroma::MInvCGAccum	(	const DiffLinearOperatorArray< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const multi1d< LatticeFermion > &	chi,
		multi1d< LatticeFermion > &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		const int	MaxCG,
		int &	n_count
	)

Definition at line 478 of file minvcg_accumulate_array.cc.

References Chroma::MInvCGAccum_a().

template<>

void Chroma::MInvCGAccum	(	const LinearOperatorArray< LatticeFermion > &	M,
		const multi1d< LatticeFermion > &	chi,
		multi1d< LatticeFermion > &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		const int	MaxCG,
		int &	n_count
	)

Definition at line 462 of file minvcg_accumulate_array.cc.

References Chroma::MInvCGAccum_a().

Referenced by Chroma::MdagMMultiSysSolverCGAccumulateArray< T >::operator()().

template<typename T, typename C>

void Chroma::MInvCGAccum_a	(	const C &	A,
		const multi1d< T > &	chi,
		multi1d< T > &	psi,
		const Real &	norm,
		const multi1d< Real > &	residues,
		const multi1d< Real > &	poles,
		const Real &	RsdCG,
		const int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 75 of file minvcg_accumulate_array.cc.

References Chroma::chi_norm, END_CODE, Chroma::PLUS, and START_CODE.

Referenced by Chroma::MInvCGAccum().

template<typename T, typename P, typename Q>

void Chroma::MInvMR	(	const DiffLinearOperator< T, P, Q > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

template<typename T>

void Chroma::MInvMR	(	const LinearOperator< T > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

template<>

void Chroma::MInvMR	(	const DiffLinearOperator< LatticeFermion, multi1d< LatticeColorMatrix >, multi1d< LatticeColorMatrix > > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdMR,
		int	MaxMR,
		int &	n_count
	)

Definition at line 273 of file minvmr.cc.

References Chroma::MInvMR_a().

template<>

void Chroma::MInvMR	(	const LinearOperator< LatticeFermion > &	M,
		const LatticeFermion &	chi,
		multi1d< LatticeFermion > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdMR,
		int	MaxMR,
		int &	n_count
	)

Definition at line 259 of file minvmr.cc.

References Chroma::MInvMR_a().

Referenced by Chroma::OverlapFermActBase::multiQprop(), and Chroma::LinOpMultiSysSolverMR< T >::operator()().

template<typename T>

void Chroma::MInvMR_a	(	const LinearOperator< T > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdMR,
		int	MaxMR,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 75 of file minvmr.cc.

References Chroma::chi_norm, END_CODE, Chroma::GramSchm(), Chroma::Ncb, Chroma::LinearOperator< T >::nFlops(), NOperEig, Chroma::PLUS, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::MInvMR().

template<typename T>

void Chroma::MInvRelCG_a	(	const LinearOperator< T > &	A,
		const T &	chi,
		multi1d< T > &	psi,
		const multi1d< Real > &	shifts,
		const multi1d< Real > &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Multishift Conjugate-Gradient (CG1) algorithm for a Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations Method used is described in Jegerlehner, hep-lat/9708029

We are searching in a subspace orthogonal to the eigenvectors EigVec of A. The source chi is assumed to already be orthogonal!

Chi = A . Psi

Algorithm:

Arguments:

A Hermitian linear operator (Read) Chi Source (Read) Psi array of solutions (Write) shifts shifts of form A + mass (Read) RsdCG residual accuracy (Read/Write) n_count Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Ap Temporary for M.p

MaxCG Maximum number of CG iterations allowed

Subroutines: A Apply matrix hermitian A to vector

Definition at line 71 of file minv_rel_cg.cc.

References Chroma::chi_norm, END_CODE, Chroma::PLUS, START_CODE, Chroma::LinearOperator< T >::subset(), and T.

Referenced by Chroma::MInvRelCG().

void Chroma::normGramSchmidt	(	multi2d< LatticeFermionD > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Definition at line 123 of file norm_gram_schm.cc.

References Chroma::normGramSchmidtArray_T().

void Chroma::normGramSchmidt	(	multi2d< LatticeFermionF > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Definition at line 115 of file norm_gram_schm.cc.

References Chroma::normGramSchmidtArray_T().

void Chroma::normGramSchmidt	(	multi1d< LatticeStaggeredFermionD > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Definition at line 105 of file norm_gram_schm.cc.

void Chroma::normGramSchmidt	(	multi1d< LatticeStaggeredFermionF > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Definition at line 97 of file norm_gram_schm.cc.

void Chroma::normGramSchmidt	(	multi1d< LatticeFermionD > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Definition at line 89 of file norm_gram_schm.cc.

void Chroma::normGramSchmidt	(	multi1d< LatticeFermionF > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Gram-Schmidt with normalization.

Definition at line 81 of file norm_gram_schm.cc.

Referenced by Chroma::InvEigCG2Env::InvEigCG2_T(), Chroma::InvEigCG2ArrayEnv::old_InvEigCG2_T(), and Chroma::sysSolver().

template<typename T>

void Chroma::normGramSchmidt_T	(	multi1d< T > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Gram-Schmidt with normalization.

Definition at line 15 of file norm_gram_schm.cc.

References Chroma::in.

template<typename T>

void Chroma::normGramSchmidtArray_T	(	multi2d< T > &	vec,
		int	f,
		int	t,
		const Subset &	sub
	)

Gram-Schmidt with normalization.

Definition at line 46 of file norm_gram_schm.cc.

References Chroma::in.

Referenced by Chroma::normGramSchmidt().

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverOptEigCGParams &	param
	)

Definition at line 37 of file syssolver_OPTeigcg_params.cc.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverOptEigBiCGParams &	param
	)

Definition at line 37 of file syssolver_OPTeigbicg_params.cc.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverMRParams &	param
	)

Definition at line 12 of file syssolver_mr_params.cc.

References Chroma::SysSolverMRParams::MaxMR, Chroma::SysSolverMRParams::MROver, Chroma::read(), and Chroma::SysSolverMRParams::RsdMR.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverEigCGParams &	param
	)

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverCGParams &	param
	)

Definition at line 12 of file syssolver_cg_params.cc.

References Chroma::SysSolverCGParams::MaxCG, Chroma::SysSolverCGParams::MaxCGRestart, Chroma::SysSolverCGParams::MinCG, Chroma::read(), Chroma::SysSolverCGParams::RsdCG, and Chroma::SysSolverCGParams::RsdCGRestart.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		SysSolverBiCGStabParams &	param
	)

Definition at line 12 of file syssolver_bicgstab_params.cc.

References Chroma::SysSolverBiCGStabParams::MaxBiCGStab, Chroma::read(), and Chroma::SysSolverBiCGStabParams::RsdBiCGStab.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		MultiSysSolverMRParams &	param
	)

Definition at line 12 of file multi_syssolver_mr_params.cc.

References Chroma::MultiSysSolverMRParams::MaxCG, Chroma::read(), and Chroma::MultiSysSolverMRParams::RsdCG.

void Chroma::read	(	XMLReader &	xml,
		const string &	path,
		MultiSysSolverCGParams &	param
	)

Definition at line 12 of file multi_syssolver_cg_params.cc.

References Chroma::MultiSysSolverCGParams::MaxCG, Chroma::read(), and Chroma::MultiSysSolverCGParams::RsdCG.

void WlInvCG2	(	const LinearOperator &	M,
		const LatticeFermion &	chi,
		LatticeFermion &	psi,
		const Real &	RsdCG,
		int	MaxCG,
		int &	n_count
	)

Conjugate-Gradient (CGNE) algorithm for a generic Linear Operator.

This subroutine uses the Conjugate Gradient (CG) algorithm to find the solution of the set of linear equations

Chi = A . Psi

where A = M^dag . M

Algorithm:

Arguments:

Parameters:

	M	Linear Operator (Read)
	chi	Source (Read)
	psi	Solution (Modify)
	RsdCG	CG residual accuracy (Read)
	MaxCG	Maximum CG iterations (Read)
	n_count	Number of CG iteration (Write)

Local Variables:

p Direction vector r Residual vector cp | r[k] |**2 c | r[k-1] |**2 k CG iteration counter a a[k] b b[k+1] d < p[k], A.p[k] > Mp Temporary for M.p

Subroutines: + A Apply matrix M or M to vector

Operations:

2 A + 2 Nc Ns + N_Count ( 2 A + 10 Nc Ns )

Definition at line 70 of file t_wlinvcg.cc.

References Chroma::MINUS, Chroma::PLUS, and Chroma::LinearOperator< T >::subset().

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverOptEigCGParams &	param
	)

Definition at line 83 of file syssolver_OPTeigcg_params.cc.

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverOptEigBiCGParams &	param
	)

Definition at line 89 of file syssolver_OPTeigbicg_params.cc.

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverMRParams &	param
	)

Definition at line 26 of file syssolver_mr_params.cc.

References Chroma::SysSolverMRParams::MaxMR, Chroma::SysSolverMRParams::MROver, Chroma::SysSolverMRParams::RsdMR, and Chroma::write().

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverEigCGParams &	param
	)

Definition at line 89 of file syssolver_eigcg_params.cc.

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverCGParams &	param
	)

Definition at line 46 of file syssolver_cg_params.cc.

References Chroma::SysSolverCGParams::MaxCG, Chroma::SysSolverCGParams::MaxCGRestart, Chroma::SysSolverCGParams::MinCG, Chroma::SysSolverCGParams::RsdCG, Chroma::SysSolverCGParams::RsdCGRestart, and Chroma::write().

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const SysSolverBiCGStabParams &	param
	)

Definition at line 22 of file syssolver_bicgstab_params.cc.

References Chroma::SysSolverBiCGStabParams::MaxBiCGStab, Chroma::SysSolverBiCGStabParams::RsdBiCGStab, and Chroma::write().

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const MultiSysSolverMRParams &	param
	)

Definition at line 21 of file multi_syssolver_mr_params.cc.

References Chroma::MultiSysSolverMRParams::MaxCG, Chroma::MultiSysSolverMRParams::RsdCG, and Chroma::write().

void Chroma::write	(	XMLWriter &	xml,
		const string &	path,
		const MultiSysSolverCGParams &	param
	)

Definition at line 21 of file multi_syssolver_cg_params.cc.

References Chroma::MultiSysSolverCGParams::MaxCG, Chroma::MultiSysSolverCGParams::RsdCG, and Chroma::write().

Generated on Sun Nov 22 04:38:37 2009 for CHROMA by

1.4.7

Inverters [Fermion action support]

Namespaces

Classes

Typedefs

Functions

Detailed Description

Typedef Documentation

Function Documentation

Inverters
[Fermion action support]