portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 15:55:37 +00:00
Code Releases Activity
848104b1a9
Grid/lib/qcd/action/fermion/PartialFractionFermion5D.h

70 lines
2.5 KiB
C
Raw Normal View History

Domain wall fermions now invert ; have the basis set up for Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson) Approx Representation Kernel. All are done with space-time taking part in checkerboarding, Ls uncheckerboarded Have only so far tested the Domain Wall limit of mobius, and at that only checked that it i) Inverts ii) 5dim DW == Ls copies of 4dim D2 iii) MeeInv Mee == 1 iv) Meo+Mee+Moe+Moo == M unprec. v) MpcDagMpc is hermitan vi) Mdag is the adjoint of M between stochastic vectors. That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
#ifndef GRID_QCD_PARTIAL_FRACTION_H
#define GRID_QCD_PARTIAL_FRACTION_H
namespace Grid {
namespace QCD {
class PartialFractionFermion5D : public WilsonFermion5D
{
public:
PartialFraction Hw with Zolo and Tanh approx converged under CG and passed EO breakdown and hermiticity tests.
2015-06-04 13:28:37 +01:00
const int part_frac_chroma_convention=1;
void Meooe_internal(const LatticeFermion &in, LatticeFermion &out,int dag);
void Mooee_internal(const LatticeFermion &in, LatticeFermion &out,int dag);
void MooeeInv_internal(const LatticeFermion &in, LatticeFermion &out,int dag);
void M_internal(const LatticeFermion &in, LatticeFermion &out,int dag);
Domain wall fermions now invert ; have the basis set up for Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson) Approx Representation Kernel. All are done with space-time taking part in checkerboarding, Ls uncheckerboarded Have only so far tested the Domain Wall limit of mobius, and at that only checked that it i) Inverts ii) 5dim DW == Ls copies of 4dim D2 iii) MeeInv Mee == 1 iv) Meo+Mee+Moe+Moo == M unprec. v) MpcDagMpc is hermitan vi) Mdag is the adjoint of M between stochastic vectors. That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
// override multiply
virtual RealD M (const LatticeFermion &in, LatticeFermion &out);
virtual RealD Mdag (const LatticeFermion &in, LatticeFermion &out);
// half checkerboard operaions
virtual void Meooe (const LatticeFermion &in, LatticeFermion &out);
virtual void MeooeDag (const LatticeFermion &in, LatticeFermion &out);
virtual void Mooee (const LatticeFermion &in, LatticeFermion &out);
virtual void MooeeDag (const LatticeFermion &in, LatticeFermion &out);
virtual void MooeeInv (const LatticeFermion &in, LatticeFermion &out);
virtual void MooeeInvDag (const LatticeFermion &in, LatticeFermion &out);
Jackson smoothed chebyshev and (untested) completion of force terms for Cayley, Partial and Cont fraction dwf and overlap. have even odd and unprec forces.
2015-07-31 21:58:35 +01:00
// force terms; five routines; default to Dhop on diagonal
virtual void MDeriv (LatticeGaugeField &mat,const LatticeFermion &U,const LatticeFermion &V,int dag);
virtual void MoeDeriv(LatticeGaugeField &mat,const LatticeFermion &U,const LatticeFermion &V,int dag);
virtual void MeoDeriv(LatticeGaugeField &mat,const LatticeFermion &U,const LatticeFermion &V,int dag);
PartialFraction Hw with Zolo and Tanh approx converged under CG and passed EO breakdown and hermiticity tests.
2015-06-04 13:28:37 +01:00
virtual void Instantiatable(void) =0; // ensure no make-eee
Domain wall fermions now invert ; have the basis set up for Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson) Approx Representation Kernel. All are done with space-time taking part in checkerboarding, Ls uncheckerboarded Have only so far tested the Domain Wall limit of mobius, and at that only checked that it i) Inverts ii) 5dim DW == Ls copies of 4dim D2 iii) MeeInv Mee == 1 iv) Meo+Mee+Moe+Moo == M unprec. v) MpcDagMpc is hermitan vi) Mdag is the adjoint of M between stochastic vectors. That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
Conjugate residual algorithm; some more unary functions
2015-06-08 12:04:59 +01:00
// Efficient support for multigrid coarsening
virtual void Mdir (const LatticeFermion &in, LatticeFermion &out,int dir,int disp);
Domain wall fermions now invert ; have the basis set up for Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson) Approx Representation Kernel. All are done with space-time taking part in checkerboarding, Ls uncheckerboarded Have only so far tested the Domain Wall limit of mobius, and at that only checked that it i) Inverts ii) 5dim DW == Ls copies of 4dim D2 iii) MeeInv Mee == 1 iv) Meo+Mee+Moe+Moo == M unprec. v) MpcDagMpc is hermitan vi) Mdag is the adjoint of M between stochastic vectors. That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
// Constructors
PartialFractionFermion5D(LatticeGaugeField &_Umu,
GridCartesian &FiveDimGrid,
GridRedBlackCartesian &FiveDimRedBlackGrid,
GridCartesian &FourDimGrid,
GridRedBlackCartesian &FourDimRedBlackGrid,
RealD _mass,RealD M5);
PartialFraction Hw with Zolo and Tanh approx converged under CG and passed EO breakdown and hermiticity tests.
2015-06-04 13:28:37 +01:00
protected:
virtual void SetCoefficientsTanh(Approx::zolotarev_data *zdata,RealD scale);
virtual void SetCoefficientsZolotarev(RealD zolo_hi,Approx::zolotarev_data *zdata);
// Part frac
RealD mass;
RealD dw_diag;
RealD R;
RealD amax;
RealD scale;
std::vector<double> p;
std::vector<double> q;
Domain wall fermions now invert ; have the basis set up for Tanh/Zolo * (Cayley/PartFrac/ContFrac) * (Mobius/Shamir/Wilson) Approx Representation Kernel. All are done with space-time taking part in checkerboarding, Ls uncheckerboarded Have only so far tested the Domain Wall limit of mobius, and at that only checked that it i) Inverts ii) 5dim DW == Ls copies of 4dim D2 iii) MeeInv Mee == 1 iv) Meo+Mee+Moe+Moo == M unprec. v) MpcDagMpc is hermitan vi) Mdag is the adjoint of M between stochastic vectors. That said, the RB schur solve, RB MpcDagMpc solve, Unprec solve all converge and the true residual becomes small; so pretty good tests.
2015-06-02 16:57:12 +01:00
};
}
}
#endif
Copy Permalink