1
0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-09 23:45:36 +00:00
Grid/tests/hmc/Test_action_dwf_gparity2fvs1f.cc

258 lines
9.4 KiB
C++
Raw Permalink Normal View History

Imported changes from feature/gparity_HMC branch: Added a bounds-check function for the RHMC with arbitrary power Added a pseudofermion action for the rational ratio with an arbitrary power and a mixed-precision variant of the same. The existing one-flavor rational ratio class now uses the general class under the hood To support testing of the two-flavor even-odd ratio pseudofermion, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a new HMC runner start type: CheckpointStartReseed, which reseeds the RNG from scratch, allowing for the creation of new evolution streams from an existing checkpoint. Added log output of seeds used when the RNG is seeded. EOFA changes: To support mixed-precision inversion, generalized the class to maintain a separate solver for the L and R operators in the heatbath (separate solvers are already implemented for the other stages) To support mixed-precision, the action of setting the operator shift coefficients is now maintained in a virtual function. A derived class for mixed-precision solvers ensures the coefficients are applied to both the double and single-prec operators The ||^2 of the random source is now stored by the heatbath and compared to the initial action when it is computed. These should be equal but may differ if the rational bounds are not chosen correctly, hence serving as a useful and free test Fixed calculation of M_eofa (previously incomplete and #if'd out) Added functionality to compute M_eofa^-1 to complement the calculation of M_eofa (both are equally expensive!) To support testing, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a test program which computes the G-parity force using the 1 and 2 flavor implementations and compares the result. Test supports DWF, EOFA and DSDR actions, chosen by a command line option. The Mobius EOFA force test now also checks the rational approximation used for the heatbath Added a test program for the mixed precision EOFA compared to the double-prec implementation, G-parity HMC test now applied GPBC in the y direction and not the t direction (GPBC in t are no longer supported) and checkpoints after every configuration Added a test program which computes the two-flavor G-parity action (via RHMC) with both the 1 and 2 flavor implementations and checks they agree Added a test program to check the implementation of M_eofa^{-1}
2022-06-22 15:27:48 +01:00
/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: tests/hmc/Test_action_dwf_gparity2fvs1f.cc
Copyright (C) 2015
Author: Christopher Kelly <ckelly@bnl.gov>
Author: paboyle <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 */
#include <Grid/Grid.h>
using namespace Grid;
template<typename FermionField2f, typename FermionField1f>
void copy2fTo1fFermionField(FermionField1f &out, const FermionField2f &in, int gpdir){
auto f0_halfgrid = PeekIndex<GparityFlavourIndex>(in,0); //on 2f Grid
FermionField1f f0_fullgrid_dbl(out.Grid());
Replicate(f0_halfgrid, f0_fullgrid_dbl); //double it up to live on the 1f Grid
auto f1_halfgrid = PeekIndex<GparityFlavourIndex>(in,1);
FermionField1f f1_fullgrid_dbl(out.Grid());
Replicate(f1_halfgrid, f1_fullgrid_dbl);
const Coordinate &dim_2f = in.Grid()->GlobalDimensions();
const Coordinate &dim_1f = out.Grid()->GlobalDimensions();
//We have to be careful for 5d fields; the s-direction is placed before the x,y,z,t and so we need to shift gpdir by 1
std::cout << "gpdir " << gpdir << std::endl;
gpdir+=1;
std::cout << "gpdir for 5D fields " << gpdir << std::endl;
std::cout << "dim_2f " << dim_2f << std::endl;
std::cout << "dim_1f " << dim_1f << std::endl;
assert(dim_1f[gpdir] == 2*dim_2f[gpdir]);
LatticeInteger xcoor_1f(out.Grid()); //5d lattice integer
LatticeCoordinate(xcoor_1f,gpdir);
2022-11-16 03:58:30 +00:00
Integer L = dim_2f[gpdir];
Imported changes from feature/gparity_HMC branch: Added a bounds-check function for the RHMC with arbitrary power Added a pseudofermion action for the rational ratio with an arbitrary power and a mixed-precision variant of the same. The existing one-flavor rational ratio class now uses the general class under the hood To support testing of the two-flavor even-odd ratio pseudofermion, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a new HMC runner start type: CheckpointStartReseed, which reseeds the RNG from scratch, allowing for the creation of new evolution streams from an existing checkpoint. Added log output of seeds used when the RNG is seeded. EOFA changes: To support mixed-precision inversion, generalized the class to maintain a separate solver for the L and R operators in the heatbath (separate solvers are already implemented for the other stages) To support mixed-precision, the action of setting the operator shift coefficients is now maintained in a virtual function. A derived class for mixed-precision solvers ensures the coefficients are applied to both the double and single-prec operators The ||^2 of the random source is now stored by the heatbath and compared to the initial action when it is computed. These should be equal but may differ if the rational bounds are not chosen correctly, hence serving as a useful and free test Fixed calculation of M_eofa (previously incomplete and #if'd out) Added functionality to compute M_eofa^-1 to complement the calculation of M_eofa (both are equally expensive!) To support testing, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a test program which computes the G-parity force using the 1 and 2 flavor implementations and compares the result. Test supports DWF, EOFA and DSDR actions, chosen by a command line option. The Mobius EOFA force test now also checks the rational approximation used for the heatbath Added a test program for the mixed precision EOFA compared to the double-prec implementation, G-parity HMC test now applied GPBC in the y direction and not the t direction (GPBC in t are no longer supported) and checkpoints after every configuration Added a test program which computes the two-flavor G-parity action (via RHMC) with both the 1 and 2 flavor implementations and checks they agree Added a test program to check the implementation of M_eofa^{-1}
2022-06-22 15:27:48 +01:00
out = where(xcoor_1f < L, f0_fullgrid_dbl, f1_fullgrid_dbl);
}
//Both have the same field type
void copy2fTo1fGaugeField(LatticeGaugeField &out, const LatticeGaugeField &in, int gpdir){
LatticeGaugeField U_dbl(out.Grid());
Replicate(in, U_dbl);
LatticeGaugeField Uconj_dbl = conjugate( U_dbl );
const Coordinate &dim_2f = in.Grid()->GlobalDimensions();
LatticeInteger xcoor_1f(out.Grid());
LatticeCoordinate(xcoor_1f,gpdir);
2022-11-16 03:58:30 +00:00
Integer L = dim_2f[gpdir];
Imported changes from feature/gparity_HMC branch: Added a bounds-check function for the RHMC with arbitrary power Added a pseudofermion action for the rational ratio with an arbitrary power and a mixed-precision variant of the same. The existing one-flavor rational ratio class now uses the general class under the hood To support testing of the two-flavor even-odd ratio pseudofermion, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a new HMC runner start type: CheckpointStartReseed, which reseeds the RNG from scratch, allowing for the creation of new evolution streams from an existing checkpoint. Added log output of seeds used when the RNG is seeded. EOFA changes: To support mixed-precision inversion, generalized the class to maintain a separate solver for the L and R operators in the heatbath (separate solvers are already implemented for the other stages) To support mixed-precision, the action of setting the operator shift coefficients is now maintained in a virtual function. A derived class for mixed-precision solvers ensures the coefficients are applied to both the double and single-prec operators The ||^2 of the random source is now stored by the heatbath and compared to the initial action when it is computed. These should be equal but may differ if the rational bounds are not chosen correctly, hence serving as a useful and free test Fixed calculation of M_eofa (previously incomplete and #if'd out) Added functionality to compute M_eofa^-1 to complement the calculation of M_eofa (both are equally expensive!) To support testing, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a test program which computes the G-parity force using the 1 and 2 flavor implementations and compares the result. Test supports DWF, EOFA and DSDR actions, chosen by a command line option. The Mobius EOFA force test now also checks the rational approximation used for the heatbath Added a test program for the mixed precision EOFA compared to the double-prec implementation, G-parity HMC test now applied GPBC in the y direction and not the t direction (GPBC in t are no longer supported) and checkpoints after every configuration Added a test program which computes the two-flavor G-parity action (via RHMC) with both the 1 and 2 flavor implementations and checks they agree Added a test program to check the implementation of M_eofa^{-1}
2022-06-22 15:27:48 +01:00
out = where(xcoor_1f < L, U_dbl, Uconj_dbl);
}
std::ostream & operator<<(std::ostream &os, const Coordinate &x){
os << "(";
for(int i=0;i<x.size();i++) os << x[i] << (i<x.size()-1 ? " " : "");
os << ")";
return os;
}
int main(int argc, char **argv) {
using namespace Grid;
Grid_init(&argc, &argv);
int threads = GridThread::GetThreads();
std::cout << GridLogMessage << "Grid is setup to use " << threads << " threads" << std::endl;
int Ls = 16;
Coordinate latt_2f = GridDefaultLatt();
Coordinate simd_layout = GridDefaultSimd(Nd, vComplexD::Nsimd());
Coordinate mpi_layout = GridDefaultMpi();
int mu = 0; //Gparity direction
Coordinate latt_1f = latt_2f;
latt_1f[mu] *= 2;
GridCartesian * UGrid_1f = SpaceTimeGrid::makeFourDimGrid(latt_1f, simd_layout, mpi_layout);
GridRedBlackCartesian * UrbGrid_1f = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid_1f);
GridCartesian * FGrid_1f = SpaceTimeGrid::makeFiveDimGrid(Ls,UGrid_1f);
GridRedBlackCartesian * FrbGrid_1f = SpaceTimeGrid::makeFiveDimRedBlackGrid(Ls,UGrid_1f);
GridCartesian * UGrid_2f = SpaceTimeGrid::makeFourDimGrid(latt_2f, simd_layout, mpi_layout);
GridRedBlackCartesian * UrbGrid_2f = SpaceTimeGrid::makeFourDimRedBlackGrid(UGrid_2f);
GridCartesian * FGrid_2f = SpaceTimeGrid::makeFiveDimGrid(Ls,UGrid_2f);
GridRedBlackCartesian * FrbGrid_2f = SpaceTimeGrid::makeFiveDimRedBlackGrid(Ls,UGrid_2f);
std::cout << "SIMD layout " << simd_layout << std::endl;
std::cout << "MPI layout " << mpi_layout << std::endl;
std::cout << "2f dimensions " << latt_2f << std::endl;
std::cout << "1f dimensions " << latt_1f << std::endl;
std::vector<int> seeds4({1,2,3,4});
std::vector<int> seeds5({5,6,7,8});
GridParallelRNG RNG5_2f(FGrid_2f); RNG5_2f.SeedFixedIntegers(seeds5);
GridParallelRNG RNG4_2f(UGrid_2f); RNG4_2f.SeedFixedIntegers(seeds4);
std::cout << "Generating hot 2f gauge configuration" << std::endl;
LatticeGaugeField Umu_2f(UGrid_2f);
SU<Nc>::HotConfiguration(RNG4_2f,Umu_2f);
std::cout << "Copying 2f->1f gauge field" << std::endl;
LatticeGaugeField Umu_1f(UGrid_1f);
copy2fTo1fGaugeField(Umu_1f, Umu_2f, mu);
typedef GparityWilsonImplR FermionImplPolicy2f;
2022-11-16 03:58:30 +00:00
typedef GparityDomainWallFermionD FermionAction2f;
Imported changes from feature/gparity_HMC branch: Added a bounds-check function for the RHMC with arbitrary power Added a pseudofermion action for the rational ratio with an arbitrary power and a mixed-precision variant of the same. The existing one-flavor rational ratio class now uses the general class under the hood To support testing of the two-flavor even-odd ratio pseudofermion, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a new HMC runner start type: CheckpointStartReseed, which reseeds the RNG from scratch, allowing for the creation of new evolution streams from an existing checkpoint. Added log output of seeds used when the RNG is seeded. EOFA changes: To support mixed-precision inversion, generalized the class to maintain a separate solver for the L and R operators in the heatbath (separate solvers are already implemented for the other stages) To support mixed-precision, the action of setting the operator shift coefficients is now maintained in a virtual function. A derived class for mixed-precision solvers ensures the coefficients are applied to both the double and single-prec operators The ||^2 of the random source is now stored by the heatbath and compared to the initial action when it is computed. These should be equal but may differ if the rational bounds are not chosen correctly, hence serving as a useful and free test Fixed calculation of M_eofa (previously incomplete and #if'd out) Added functionality to compute M_eofa^-1 to complement the calculation of M_eofa (both are equally expensive!) To support testing, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a test program which computes the G-parity force using the 1 and 2 flavor implementations and compares the result. Test supports DWF, EOFA and DSDR actions, chosen by a command line option. The Mobius EOFA force test now also checks the rational approximation used for the heatbath Added a test program for the mixed precision EOFA compared to the double-prec implementation, G-parity HMC test now applied GPBC in the y direction and not the t direction (GPBC in t are no longer supported) and checkpoints after every configuration Added a test program which computes the two-flavor G-parity action (via RHMC) with both the 1 and 2 flavor implementations and checks they agree Added a test program to check the implementation of M_eofa^{-1}
2022-06-22 15:27:48 +01:00
typedef typename FermionAction2f::FermionField FermionField2f;
typedef WilsonImplR FermionImplPolicy1f;
2022-11-16 03:58:30 +00:00
typedef DomainWallFermionD FermionAction1f;
Imported changes from feature/gparity_HMC branch: Added a bounds-check function for the RHMC with arbitrary power Added a pseudofermion action for the rational ratio with an arbitrary power and a mixed-precision variant of the same. The existing one-flavor rational ratio class now uses the general class under the hood To support testing of the two-flavor even-odd ratio pseudofermion, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a new HMC runner start type: CheckpointStartReseed, which reseeds the RNG from scratch, allowing for the creation of new evolution streams from an existing checkpoint. Added log output of seeds used when the RNG is seeded. EOFA changes: To support mixed-precision inversion, generalized the class to maintain a separate solver for the L and R operators in the heatbath (separate solvers are already implemented for the other stages) To support mixed-precision, the action of setting the operator shift coefficients is now maintained in a virtual function. A derived class for mixed-precision solvers ensures the coefficients are applied to both the double and single-prec operators The ||^2 of the random source is now stored by the heatbath and compared to the initial action when it is computed. These should be equal but may differ if the rational bounds are not chosen correctly, hence serving as a useful and free test Fixed calculation of M_eofa (previously incomplete and #if'd out) Added functionality to compute M_eofa^-1 to complement the calculation of M_eofa (both are equally expensive!) To support testing, separated the functionality of generating the random field and performing the heatbath step, and added a method to obtain the pseudofermion field Added a test program which computes the G-parity force using the 1 and 2 flavor implementations and compares the result. Test supports DWF, EOFA and DSDR actions, chosen by a command line option. The Mobius EOFA force test now also checks the rational approximation used for the heatbath Added a test program for the mixed precision EOFA compared to the double-prec implementation, G-parity HMC test now applied GPBC in the y direction and not the t direction (GPBC in t are no longer supported) and checkpoints after every configuration Added a test program which computes the two-flavor G-parity action (via RHMC) with both the 1 and 2 flavor implementations and checks they agree Added a test program to check the implementation of M_eofa^{-1}
2022-06-22 15:27:48 +01:00
typedef typename FermionAction1f::FermionField FermionField1f;
std::cout << "Generating eta 2f" << std::endl;
FermionField2f eta_2f(FGrid_2f);
gaussian(RNG5_2f, eta_2f);
RealD scale = std::sqrt(0.5);
eta_2f=eta_2f*scale;
std::cout << "Copying 2f->1f eta" << std::endl;
FermionField1f eta_1f(FGrid_1f);
copy2fTo1fFermionField(eta_1f, eta_2f, mu);
Real beta = 2.13;
Real light_mass = 0.01;
Real strange_mass = 0.032;
Real pv_mass = 1.0;
RealD M5 = 1.8;
//Setup the Dirac operators
std::cout << "Initializing Dirac operators" << std::endl;
FermionAction2f::ImplParams Params_2f;
Params_2f.twists[mu] = 1;
Params_2f.twists[Nd-1] = 1; //APBC in time direction
//note 'Num' and 'Den' here refer to the determinant ratio, not the operator ratio in the pseudofermion action where the two are inverted
//to my mind the Pauli Villars and 'denominator' are synonymous but the Grid convention has this as the 'Numerator' operator in the RHMC implementation
FermionAction2f NumOp_2f(Umu_2f,*FGrid_2f,*FrbGrid_2f,*UGrid_2f, *UrbGrid_2f, light_mass,M5,Params_2f);
FermionAction2f DenOp_2f(Umu_2f,*FGrid_2f,*FrbGrid_2f,*UGrid_2f, *UrbGrid_2f, pv_mass, M5,Params_2f);
FermionAction1f::ImplParams Params_1f;
Params_1f.boundary_phases[mu] = -1; //antiperiodic in doubled lattice in GP direction
Params_1f.boundary_phases[Nd-1] = -1;
FermionAction1f NumOp_1f(Umu_1f,*FGrid_1f,*FrbGrid_1f,*UGrid_1f, *UrbGrid_1f, light_mass,M5,Params_1f);
FermionAction1f DenOp_1f(Umu_1f,*FGrid_1f,*FrbGrid_1f,*UGrid_1f, *UrbGrid_1f, pv_mass, M5,Params_1f);
//Test the replication routines by running a CG on eta
double StoppingCondition = 1e-10;
double MaxCGIterations = 30000;
ConjugateGradient<FermionField2f> CG_2f(StoppingCondition,MaxCGIterations);
ConjugateGradient<FermionField1f> CG_1f(StoppingCondition,MaxCGIterations);
NumOp_1f.ImportGauge(Umu_1f);
NumOp_2f.ImportGauge(Umu_2f);
FermionField1f test_1f(FGrid_1f);
FermionField2f test_2f(FGrid_2f);
MdagMLinearOperator<FermionAction1f, FermionField1f> Linop_1f(NumOp_1f);
MdagMLinearOperator<FermionAction2f, FermionField2f> Linop_2f(NumOp_2f);
CG_1f(Linop_1f, eta_1f, test_1f);
CG_2f(Linop_2f, eta_2f, test_2f);
RealD test_1f_norm = norm2(test_1f);
RealD test_2f_norm = norm2(test_2f);
std::cout << "Verification of replication routines: " << test_1f_norm << " " << test_2f_norm << " " << test_1f_norm - test_2f_norm << std::endl;
#if 1
typedef GeneralEvenOddRatioRationalPseudoFermionAction<FermionImplPolicy2f> Action2f;
typedef GeneralEvenOddRatioRationalPseudoFermionAction<FermionImplPolicy1f> Action1f;
RationalActionParams rational_params;
rational_params.inv_pow = 2;
rational_params.lo = 1e-5;
rational_params.hi = 32;
rational_params.md_degree = 16;
rational_params.action_degree = 16;
Action2f action_2f(DenOp_2f, NumOp_2f, rational_params);
Action1f action_1f(DenOp_1f, NumOp_1f, rational_params);
#else
typedef TwoFlavourEvenOddRatioPseudoFermionAction<FermionImplPolicy2f> Action2f;
typedef TwoFlavourEvenOddRatioPseudoFermionAction<FermionImplPolicy1f> Action1f;
Action2f action_2f(DenOp_2f, NumOp_2f, CG_2f, CG_2f);
Action1f action_1f(DenOp_1f, NumOp_1f, CG_1f, CG_1f);
#endif
std::cout << "Action refresh" << std::endl;
action_2f.refresh(Umu_2f, eta_2f);
action_1f.refresh(Umu_1f, eta_1f);
std::cout << "Action compute post heatbath" << std::endl;
RealD S_2f = action_2f.S(Umu_2f);
RealD S_1f = action_1f.S(Umu_1f);
std::cout << "Action comparison post heatbath" << std::endl;
std::cout << S_2f << " " << S_1f << " " << S_2f-S_1f << std::endl;
//Change the gauge field between refresh and action eval else the matrix and inverse matrices all cancel and we just get |eta|^2
SU<Nc>::HotConfiguration(RNG4_2f,Umu_2f);
copy2fTo1fGaugeField(Umu_1f, Umu_2f, mu);
//Now compute the action with the new gauge field
std::cout << "Action compute post gauge field update" << std::endl;
S_2f = action_2f.S(Umu_2f);
S_1f = action_1f.S(Umu_1f);
std::cout << "Action comparison post gauge field update" << std::endl;
std::cout << S_2f << " " << S_1f << " " << S_2f-S_1f << std::endl;
Grid_finalize();
} // main