1
0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 15:55:37 +00:00
Grid/Hadrons/Distil.hpp
Michael Marshall f8e1941327 Implemented specialisations of NamedTensor as derived classes, however this suffers a number of problems:
1) virtual functions not available in base class constructor where I'd like to use them - e.g. IndexNames
2) Must define new constructors in derived classes
... so the specialisations are fatter than I'd like. Would prefer to revert to specifying tensor name and index name defaults in template
2019-11-08 11:55:00 +00:00

182 lines
8.0 KiB
C++

/*************************************************************************************
Grid physics library, www.github.com/paboyle/Grid
Source file: Hadrons/Distil.hpp
Copyright (C) 2015-2019
Author: Felix Erben <ferben@ed.ac.uk>
Author: Michael Marshall <Michael.Marshall@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 */
#ifndef Hadrons_Distil_hpp_
#define Hadrons_Distil_hpp_
#include <Hadrons/Global.hpp>
#include <Hadrons/EigenPack.hpp>
BEGIN_HADRONS_NAMESPACE
/******************************************************************************
NamedTensor
Eigen::Tensor of type Scalar_ and rank NumIndices_ (row-major order), together with a name for each index.
Index names are mutable, but tensor dimensionality is not (size of each dimension is mutable).
They can be persisted to / restored from disk, by default using tag Name.
During restore from disk, these validations are performed:
1) Tensor dimensionality must match
2) IndexNames are validated against current values
3) If the tensor has non-zero size, the tensor being loaded must have same extent in each dimension
******************************************************************************/
extern const std::string NamedTensorFileExtension;
template<typename Scalar_, int NumIndices_>
class NamedTensor : Serializable
{
public:
using Scalar = Scalar_;
static constexpr int NumIndices = NumIndices_;
using ET = Eigen::Tensor<Scalar_, NumIndices_, Eigen::RowMajor>;
using Index = typename ET::Index;
GRID_SERIALIZABLE_CLASS_MEMBERS(NamedTensor,
ET, tensor,
std::vector<std::string>, IndexNames );
// Name of the object and Index names as set in the constructor
const std::string &Name;
const std::vector<std::string> &DefaultIndexNames;
virtual ~NamedTensor(){};
// Default constructor (assumes tensor will be loaded from file)
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE NamedTensor(const std::string &Name_, const std::vector<std::string> &IndexNames_)
: IndexNames{IndexNames_}, Name{Name_}, DefaultIndexNames{IndexNames_} {}
// Construct a named tensor explicitly specifying size of each dimension
template<typename... IndexTypes>
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE NamedTensor(const std::string &Name_, const std::vector<std::string> &IndexNames_,
Eigen::Index firstDimension, IndexTypes... otherDimensions)
: tensor(firstDimension, otherDimensions...), IndexNames{IndexNames_}, Name{Name_}, DefaultIndexNames{IndexNames_}
{
assert(sizeof...(otherDimensions) + 1 == NumIndices_ && "NamedTensor: dimensions != tensor rank");
}
// Do my index names match the default for my type?
bool ValidateIndexNames( const std::vector<std::string> &CheckNames ) const
{
assert( CheckNames.size() == NumIndices_ && "Bug: CheckNames don't match NumIndices_" );
bool bSame{ IndexNames.size() == NumIndices_ };
for( std::size_t i = 0; bSame && i < NumIndices_; i++ )
{
bSame = IndexNames[i].size() == CheckNames[i].size()
&& std::equal( IndexNames[i].begin(), IndexNames[i].end(), CheckNames[i].begin(),
[](const char & c1, const char & c2){ return c1 == c2 || std::toupper(c1) == std::toupper(c2); });
}
return bSame;
}
bool ValidateIndexNames() const { return ValidateIndexNames(DefaultIndexNames); }
#ifdef HAVE_HDF5
using Default_Reader = Grid::Hdf5Reader;
using Default_Writer = Grid::Hdf5Writer;
#else
using Default_Reader = Grid::BinaryReader;
using Default_Writer = Grid::BinaryWriter;
#endif
void write(const std::string &FileName, const std::string &Tag) const
{
std::string FileName_{FileName};
FileName_.append( NamedTensorFileExtension );
LOG(Message) << "Writing " << Name << " to file " << FileName_ << " tag " << Tag << std::endl;
Default_Writer w( FileName_ );
write( w, Tag, *this );
}
void write(const std::string &FileName) const { return write(FileName, Name); }
// Read tensor.
// Validate:
// 1) index names (if requested)
// 2) index dimensions (if they are non-zero when called)
template<typename Reader> void read(Reader &r, bool bValidate, const std::string &Tag)
{
// Grab index names and dimensions
std::vector<std::string> OldIndexNames{std::move(IndexNames)};
const typename ET::Dimensions OldDimensions{tensor.dimensions()};
read(r, Tag, *this);
const typename ET::Dimensions & NewDimensions{tensor.dimensions()};
for (int i = 0; i < NumIndices_; i++)
assert(OldDimensions[i] == 0 || OldDimensions[i] == NewDimensions[i] && "NamedTensor::read dimension size");
if (bValidate)
assert(ValidateIndexNames(OldIndexNames) && "NamedTensor::read dimension name");
}
template<typename Reader> void read(Reader &r, bool bValidate = true) { read(r, bValidate, Name); }
inline void read (const std::string &FileName, bool bValidate, const std::string &Tag)
{
Default_Reader r(FileName + NamedTensorFileExtension);
read(r, bValidate, Tag);
}
inline void read (const std::string &FileName, bool bValidate= true) { return read(FileName, bValidate, Name); }
};
/******************************************************************************
Common elements for distillation
******************************************************************************/
BEGIN_MODULE_NAMESPACE(MDistil)
//Eigenvectors of the Laplacian
using LapEvecs = Grid::Hadrons::EigenPack<LatticeColourVector>;
// Noise vector (index order: nnoise, nt, nvec, ns)
class NoiseTensor : public NamedTensor<Complex, 4>
{
static const std::string Name_;
static const std::vector<std::string> DefaultIndexNames_;
public:
// Default constructor (assumes tensor will be loaded from file)
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE NoiseTensor() : NamedTensor{Name_, DefaultIndexNames_} {}
// Construct a named tensor explicitly specifying size of each dimension
template<typename... IndexTypes>
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE NoiseTensor(Eigen::Index nNoise, Eigen::Index nT, Eigen::Index nVec, Eigen::Index nS)
: NamedTensor{Name_, DefaultIndexNames_, nNoise, nT, nVec, nS} {}
};
class PerambTensor : public NamedTensor<SpinVector, 6>
{
static const std::string Name_;
static const std::vector<std::string> DefaultIndexNames_;
public:
// Default constructor (assumes tensor will be loaded from file)
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE PerambTensor() : NamedTensor{Name_, DefaultIndexNames_} {}
// Construct a named tensor explicitly specifying size of each dimension
template<typename... IndexTypes>
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE PerambTensor(Eigen::Index nT, Eigen::Index nVec, Eigen::Index LI, Eigen::Index nNoise, Eigen::Index nT_inv, Eigen::Index SI)
: NamedTensor{Name_, DefaultIndexNames_, nT, nVec, LI, nNoise, nT_inv, SI} {}
};
END_MODULE_NAMESPACE
END_HADRONS_NAMESPACE
#endif // Hadrons_Distil_hpp_