 |
SyTen
|
|
|
SyTen
|
|
Contains helper functions related to the product of two dense tensors, see syten::prodD() for the main entry point. More...
Classes | |
| struct | TemporaryTransposeStorage |
| Temporary transpose storage class, used during tensor-tensor products. More... | |
Functions | |
| template<Rank frank, Rank srank> | |
| void | check_dense_dims (std::array< Index, frank > const &dimA, std::array< Index, srank > const &dimB, std::array< int, frank > const &c_a, std::array< int, srank > const &c_b) |
Asserts that the dimensions of two tensors, as described by dimA and dimB, are compatible with hte product described in c_a and c_b. More... | |
| template<Rank summed, Rank frank, Rank srank, Rank rrank, typename Scalar > | |
| void | gemm_transpose (DenseTensor< frank, Scalar > const &a, DenseTensor< srank, Scalar > const &b, DenseTensor< rrank, Scalar > &r, std::array< int, frank > const &c_a, std::array< int, srank > const &c_b) |
| Product of two trivial-order tensors with subsequent transpose to obtain the desired order in the output. More... | |
| template<Rank summed, Rank frank, Rank srank, typename Scalar > | |
| DenseTensor< frank+srank - 2 *summed, Scalar > | impl (DenseTensor< frank, Scalar > const &a, DenseTensor< srank, Scalar > const &b, std::array< int, frank > const &c_a, std::array< int, srank > const &c_b, bool conjugate=false, TemporaryTransposeStorage< Scalar, frank, srank > *tts=nullptr) |
| Dense tensor-tensor contraction, default implementation. More... | |
| template<> | |
| DenseTensor< 1, HighPrec > | impl< 1 > (DenseTensor< 1, HighPrec > const &a, DenseTensor< 2, HighPrec > const &b, std::array< int, 1 > const &c_a, std::array< int, 2 > const &c_b, bool conjugate, TemporaryTransposeStorage< HighPrec, 1, 2 > *) |
| Specialisation for a matrix-vector product of HighPrec scalars. More... | |
| template<> | |
| DenseTensor< 1, double > | impl< 1 > (DenseTensor< 2, double > const &a, DenseTensor< 1, double > const &b, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b, bool conjugate, TemporaryTransposeStorage< double, 2, 1 > *) |
| Specialisation for a matrix-vector product of doubles. More... | |
| template<> | |
| DenseTensor< 1, float > | impl< 1 > (DenseTensor< 2, float > const &a, DenseTensor< 1, float > const &b, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b, bool conjugate, TemporaryTransposeStorage< float, 2, 1 > *) |
| Specialisation for a matrix-vector product of floats. More... | |
| template<> | |
| DenseTensor< 2, HighPrec > | impl< 1 > (DenseTensor< 2, HighPrec > const &a, DenseTensor< 2, HighPrec > const &b, std::array< int, 2 > const &c_a, std::array< int, 2 > const &c_b, bool conjugate, TemporaryTransposeStorage< HighPrec, 2, 2 > *) |
| Specialisation for a matrix-matrix product of HighPrec scalars. More... | |
| template<> | |
| DenseTensor< 1, std::complex< double > > | impl< 1 > (DenseTensor< 2, std::complex< double > > const &a, DenseTensor< 1, std::complex< double > > const &b, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b, bool conjugate, TemporaryTransposeStorage< std::complex< double >, 2, 1 > *) |
| Specialisation for a matrix-vector product of complex doubles. More... | |
| template<> | |
| DenseTensor< 1, std::complex< float > > | impl< 1 > (DenseTensor< 2, std::complex< float > > const &a, DenseTensor< 1, std::complex< float > > const &b, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b, bool conjugate, TemporaryTransposeStorage< std::complex< float >, 2, 1 > *) |
| Specialisation for a matrix-vector product of complex floats. More... | |
| template<Rank rank> | |
| bool | is_trivial_order_a (std::array< Index, rank > const &dims, std::array< int, rank > const &c_a) |
| Returns true if the tensor as specified by its dimensions and contraction specifier does not have to be transposed to be usable in a TTGT-based tensor transposition if it is the first tensor to be contracted. More... | |
| template<Rank rank> | |
| bool | is_trivial_order_b (std::array< Index, rank > const &dims, std::array< int, rank > const &c_b) |
| Returns true if the tensor as specified by its dimensions and contraction specifier does not have to be transposed to be usable in a TTGT-based tensor transposition if it is the second tensor to be contracted. More... | |
| template<Rank frank, Rank srank, Rank rrank> | |
| bool | is_trivial_order_r (std::array< Index, frank > const &dims_a, std::array< Index, srank > const &dims_b, std::array< int, frank > const &c_a, std::array< int, srank > const &c_b) |
Returns true if r calculated as the contraction of tensors a and b defined by their dimensions and contraction specifiers does not have to be transposed according to those specifiers. More... | |
| DenseTensor< 2, HighPrec > | mm (DenseTensor< 2, HighPrec > const &a, DenseTensor< 2, HighPrec > const &b, std::array< int, 2 > const &c_a, std::array< int, 2 > const &c_b, bool conjugate) |
| Matrix-matrix product of high-precision scalars. More... | |
| void | mm_cm (double const *at, double const *bt, double *rt, std::size_t const dx, std::size_t const dy, std::size_t const dz) |
| General column-major matrix-matrix product. More... | |
| void | mm_cm (float const *at, float const *bt, float *rt, std::size_t const dx, std::size_t const dy, std::size_t const dz) |
| General column-major matrix-matrix product. More... | |
| void | mm_cm (std::complex< double > const *at, std::complex< double > const *bt, std::complex< double > *rt, std::size_t const dx, std::size_t const dy, std::size_t const dz) |
| General column-major matrix-matrix product. More... | |
| void | mm_cm (std::complex< float > const *at, std::complex< float > const *bt, std::complex< float > *rt, std::size_t const dx, std::size_t const dy, std::size_t const dz) |
| General column-major matrix-matrix product. More... | |
| DenseTensor< 1, HighPrec > | mv (DenseTensor< 1, HighPrec > const &a, DenseTensor< 2, HighPrec > const &b, std::array< int, 1 > const &c_a, std::array< int, 2 > const &c_b, bool conjugate) |
| Matrix-vector product of high-precision scalars. More... | |
| DenseTensor< 1, double > | mv (DenseTensor< 2, double > const &a, DenseTensor< 1, double > const &v, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b) |
| Matrix-vector product of doubles. More... | |
| DenseTensor< 1, float > | mv (DenseTensor< 2, float > const &a, DenseTensor< 1, float > const &v, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b) |
| Matrix-vector product of floats. More... | |
| DenseTensor< 1, std::complex< double > > | mv (DenseTensor< 2, std::complex< double > > const &a, DenseTensor< 1, std::complex< double > > const &v, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b) |
| Matrix-vector product of std::complex<double> More... | |
| DenseTensor< 1, std::complex< float > > | mv (DenseTensor< 2, std::complex< float > > const &a, DenseTensor< 1, std::complex< float > > const &v, std::array< int, 2 > const &c_a, std::array< int, 1 > const &c_b) |
| Matrix-vector product of std::complex<float> More... | |
| template<Rank summed, Rank frank, Rank srank, typename Scalar > | |
| DenseTensor< frank+srank - 2 *summed, Scalar > | one_times_many (DenseTensor< frank, Scalar > const &a, DenseTensor< srank, Scalar > const &b, std::array< int, frank > const &c_a, std::array< int, srank > const &c_b, bool conjugate_a, bool conjugate_b) |
| Product of a singleton tensor and a larger tensor. More... | |
| template<Rank summed, Rank frank, Rank srank, typename Scalar > | |
| DenseTensor< frank+srank - 2 *summed, Scalar > | one_times_one (DenseTensor< frank, Scalar > const &a, DenseTensor< srank, Scalar > const &b, std::array< int, frank > const &, std::array< int, srank > const &, bool conjugate) |
| Product of two singleton tensors. More... | |
| template<std::size_t rank> | |
| std::vector< int > | reverse_dims_to_vec (std::array< Index, rank > const &in) |
Returns a vector with the entry of in in reversed order. More... | |
| template<Rank summed, Rank frank, typename Scalar > | |
| void | transpose_a (DenseTensor< frank, Scalar > const &a, std::array< int, frank > const &c_a, DenseTensor< frank, Scalar > &out_a, std::array< int, frank > &out_c_a) |
| Transposition of a first/a tensor to trivial order. More... | |
| template<Rank summed, Rank srank, typename Scalar > | |
| void | transpose_b (DenseTensor< srank, Scalar > const &b, std::array< int, srank > const &c_b, DenseTensor< srank, Scalar > &out_b, std::array< int, srank > &out_c_b, Conj do_conj=Conj::n()) |
| Transposition of a second/b tensor to trivial order. More... | |
| template<Rank frank, Rank srank> | |
| void | validateProductSpecifier (std::array< int, frank > const &c_a, std::array< int, srank > const &c_b, int summed=0) |
| Validates a pair of product specifiers. More... | |
Contains helper functions related to the product of two dense tensors, see syten::prodD() for the main entry point.
1.9.4