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
ff71a8e847
Grid/lib/qcd/action/fermion/CayleyFermion5D.h

156 lines
6.0 KiB
C
Raw Normal View History

Global edit adding copyright and license info to every source file.
2016-01-02 14:51:32 +00:00
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: ./lib/qcd/action/fermion/CayleyFermion5D.h
Copyright (C) 2015
Author: Peter Boyle <pabobyle@ph.ed.ac.uk>
Author: Peter Boyle <paboyle@ph.ed.ac.uk>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
See the full license in the file "LICENSE" in the top level distribution directory
*************************************************************************************/
/* END LEGAL */
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_CAYLEY_FERMION_H
#define GRID_QCD_CAYLEY_FERMION_H
namespace Grid {
namespace QCD {
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
template<class Impl>
class CayleyFermion5D : public WilsonFermion5D<Impl>
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
{
public:
* Finished the template/policy style introduction of gparity, except the gparity force terms. So valence sector looks ok. FermionOperatorImpl.h provides the policy classes. Expect HMC will introduce a smearing policy and a fermion representation change policy template param. Will also probably need multi-precision work. * HMC is running even-odd and non-checkerboarded (checked 4^4 wilson fermion/wilson gauge). There appears to be a bug in the multi-level integrator -- <e-dH> passes with single level but not with multi-level. In any case there looks to be quite a bit to clean up. This is the "const det" style implementation that is not appropriate yet for clover since it assumes that Mee is indept of the gauge fields. Easily fixed in future.
2015-08-15 23:25:49 +01:00
INHERIT_IMPL_TYPES(Impl);
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
public:
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
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
virtual RealD M (const FermionField &in, FermionField &out);
virtual RealD Mdag (const FermionField &in, FermionField &out);
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
// half checkerboard operations
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
virtual void Meooe (const FermionField &in, FermionField &out);
virtual void MeooeDag (const FermionField &in, FermionField &out);
virtual void Mooee (const FermionField &in, FermionField &out);
virtual void MooeeDag (const FermionField &in, FermionField &out);
virtual void MooeeInv (const FermionField &in, FermionField &out);
virtual void MooeeInvDag (const FermionField &in, FermionField &out);
Multiple implementations for the 5d hopping terms, depending on cache friendly ops and/or the 5th direction being vectorised All use 4d redblack.
2016-07-15 00:00:09 +01:00
virtual void Meo5D (const FermionField &psi, FermionField &chi);
virtual void M5D (const FermionField &psi, FermionField &chi);
virtual void M5Ddag(const FermionField &psi, FermionField &chi);
Dminus
2016-10-10 23:43:19 +01:00
virtual void Dminus(const FermionField &psi, FermionField &chi);
virtual void DminusDag(const FermionField &psi, FermionField &chi);
Multiple implementations for the 5d hopping terms, depending on cache friendly ops and/or the 5th direction being vectorised All use 4d redblack.
2016-07-15 00:00:09 +01:00
/////////////////////////////////////////////////////
// Instantiate different versions depending on Impl
/////////////////////////////////////////////////////
void M5D(const FermionField &psi,
const FermionField &phi,
FermionField &chi,
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> &lower,
std::vector<Coeff_t> &diag,
std::vector<Coeff_t> &upper);
Multiple implementations for the 5d hopping terms, depending on cache friendly ops and/or the 5th direction being vectorised All use 4d redblack.
2016-07-15 00:00:09 +01:00
void M5Ddag(const FermionField &psi,
const FermionField &phi,
FermionField &chi,
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> &lower,
std::vector<Coeff_t> &diag,
std::vector<Coeff_t> &upper);
Multiple implementations for the 5d hopping terms, depending on cache friendly ops and/or the 5th direction being vectorised All use 4d redblack.
2016-07-15 00:00:09 +01:00
void MooeeInternal(const FermionField &in, FermionField &out,int dag,int inv);
First pass at continued fraction; solver and even odd decomposition tests pass. Have to make ContFrac class virtual and derive end non-abstract actions for the particular cases.
2015-06-04 00:00:45 +01:00
virtual void Instantiatable(void)=0;
Conjugate residual algorithm; some more unary functions
2015-06-08 12:04:59 +01:00
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
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
virtual void MDeriv (GaugeField &mat,const FermionField &U,const FermionField &V,int dag);
virtual void MoeDeriv(GaugeField &mat,const FermionField &U,const FermionField &V,int dag);
virtual void MeoDeriv(GaugeField &mat,const FermionField &U,const FermionField &V,int dag);
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
Conjugate residual algorithm; some more unary functions
2015-06-08 12:04:59 +01:00
// Efficient support for multigrid coarsening
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
virtual void Mdir (const FermionField &in, FermionField &out,int dir,int disp);
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
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
void Meooe5D (const FermionField &in, FermionField &out);
void MeooeDag5D (const FermionField &in, FermionField &out);
Conjugate residual algorithm; some more unary functions
2015-06-08 12:04:59 +01:00
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
// protected:
RealD mass;
Mobius Caley form, Mobius Zolotarev operators. Pass Even Odd vs unprec test and hermiticity checks in tests/Grid_any_evenodd.cc; will work on inversion tests shortly.
2015-06-03 09:36:26 +01:00
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
// Cayley form Moebius (tanh and zolotarev)
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> omega;
std::vector<Coeff_t> bs; // S dependent coeffs
std::vector<Coeff_t> cs;
std::vector<Coeff_t> as;
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
// For preconditioning Cayley form
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> bee;
std::vector<Coeff_t> cee;
std::vector<Coeff_t> aee;
std::vector<Coeff_t> beo;
std::vector<Coeff_t> ceo;
std::vector<Coeff_t> aeo;
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
// LDU factorisation of the eeoo matrix
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> lee;
std::vector<Coeff_t> leem;
std::vector<Coeff_t> uee;
std::vector<Coeff_t> ueem;
std::vector<Coeff_t> dee;
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
Rework/global edit to enforce type templating of fermion operators. Allows multi-precision work and paves the way for alternate BC's and such like allowing for example G-parity which is important for K pipi programme. In particular, can drive an extra flavour index into the fermion fields using template types.
2015-08-10 20:47:44 +01:00
CayleyFermion5D(GaugeField &_Umu,
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
GridCartesian &FiveDimGrid,
GridRedBlackCartesian &FiveDimRedBlackGrid,
GridCartesian &FourDimGrid,
GridRedBlackCartesian &FourDimRedBlackGrid,
* Finished the template/policy style introduction of gparity, except the gparity force terms. So valence sector looks ok. FermionOperatorImpl.h provides the policy classes. Expect HMC will introduce a smearing policy and a fermion representation change policy template param. Will also probably need multi-precision work. * HMC is running even-odd and non-checkerboarded (checked 4^4 wilson fermion/wilson gauge). There appears to be a bug in the multi-level integrator -- <e-dH> passes with single level but not with multi-level. In any case there looks to be quite a bit to clean up. This is the "const det" style implementation that is not appropriate yet for clover since it assumes that Mee is indept of the gauge fields. Easily fixed in future.
2015-08-15 23:25:49 +01:00
RealD _mass,RealD _M5,const ImplParams &p= ImplParams());
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
Serialisation in malloc fixed
2016-11-29 22:27:55 +00:00
void CayleyReport(void);
void CayleyZeroCounters(void);
double M5Dflops;
double M5Dcalls;
double M5Dtime;
double MooeeInvFlops;
double MooeeInvCalls;
double MooeeInvTime;
Dminus
2016-10-10 23:43:19 +01:00
Mobius Caley form, Mobius Zolotarev operators. Pass Even Odd vs unprec test and hermiticity checks in tests/Grid_any_evenodd.cc; will work on inversion tests shortly.
2015-06-03 09:36:26 +01:00
protected:
First pass at continued fraction; solver and even odd decomposition tests pass. Have to make ContFrac class virtual and derive end non-abstract actions for the particular cases.
2015-06-04 00:00:45 +01:00
void SetCoefficientsZolotarev(RealD zolohi,Approx::zolotarev_data *zdata,RealD b,RealD c);
void SetCoefficientsTanh(Approx::zolotarev_data *zdata,RealD b,RealD c);
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
void SetCoefficientsInternal(RealD zolo_hi,std::vector<Coeff_t> & gamma,RealD b,RealD c);
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
};
}
}
Solving the instantiation no compile issue
2016-07-15 17:19:44 +01:00
#define INSTANTIATE_DPERP(A)\
template void CayleyFermion5D< A >::M5D(const FermionField &psi,const FermionField &phi,FermionField &chi,\
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> &lower,std::vector<Coeff_t> &diag,std::vector<Coeff_t> &upper); \
Solving the instantiation no compile issue
2016-07-15 17:19:44 +01:00
template void CayleyFermion5D< A >::M5Ddag(const FermionField &psi,const FermionField &phi,FermionField &chi,\
Feature for z-Mobius prep
2016-08-15 22:31:29 +01:00
std::vector<Coeff_t> &lower,std::vector<Coeff_t> &diag,std::vector<Coeff_t> &upper); \
Solving the instantiation no compile issue
2016-07-15 17:19:44 +01:00
template void CayleyFermion5D< A >::MooeeInv (const FermionField &psi, FermionField &chi); \
template void CayleyFermion5D< A >::MooeeInvDag (const FermionField &psi, FermionField &chi);
#define CAYLEY_DPERP_CACHE
#undef CAYLEY_DPERP_LINALG
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