Derived Types

Alias Sampler class

This class implements the Walker-Vose Alias Sampling method.

A class for tokenized reading of lines, that can be attached to open file handles.

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

@brief Re-sizeable array type that can be filled elementwise to build up a contiguous data chunk that can then be dumped to an allocatable

A class to represent information storage demand

this abstract excitation generator covers all ab initio Hamiltonians in the typical sense (i.e. up to double excitations)

Implementation for densely stored 6-index objects

This represents a distinct double excitation (compare Table 4.3 of [@DobrautzThesis2019]) A distinct double excitation consists of the two double excitations $e_{ij,kl}$ which are connected by exchange $e_{il,kj}$.

Abstract base class for excitations.

Represents the orbital indices of a 0-order excitation The array is sorted like: [srcs, tgts]

Represents the orbital indices of a 1-order excitation The array is sorted like: [srcs, tgts]

Represents the orbital indices of a 2-order excitation The array is sorted like: [srcs, tgts]

Represents the orbital indices of a 3-order excitation The array is sorted like: [srcs, tgts]

Represents an excitation with so many different indices, it has to be zero

An abstract class that supports tokenized reading of lines.

The heath bath GAS on-the-fly excitation generator

The heath bath GAS on-the-fly excitation generator

The precomputed GAS uniform excitation generator

Speficies the GAS spaces.

An index in the format of the rest of NECI [i, j, a, b] corresponds to $e_{aibj}$

An index in the format of the Shavit papers [@Hinze1981] $e_{ijkl}$

Handler for reading hdf5 tcdump files. Calls the read_op_hdf5 of the calling lMat_t

Abstract base class for lMat_t objects (6-index integrals)

A class for tokenized reading of lines, that manages the file access.

The PropVec PCHB excitation generator for doubles

PropVec PCHB excitation generator for doubles using spin orbitals, with “fast weighting” This means we choose holes via $p(AB|IJ)$

The PropVec PCHB excitation generator for doubles using spin orbitals and doing full weighting. This means that first a hole is chosen via ( p( A | I J) |{A \notin D_i} ) !! then a second hole is chosen via ( p( B | I J A) | ).

The precomputed PropVec uniform excitation generator

The precomputed PropVec uniform excitation generator

Shared memory 2-D array template with non-uniform 2nd dimension (“ragged”) of type bool

Shared memory 2-D array template with non-uniform 2nd dimension (“ragged”) of type cmplx

Shared memory 2-D array template with non-uniform 2nd dimension (“ragged”) of type int32

Shared memory 2-D array template with non-uniform 2nd dimension (“ragged”) of type int64

Shared memory 2-D array template with non-uniform 2nd dimension (“ragged”) of type real

The shared read-only hash table stores a given number of arbitrary input values in one contiguous array and addresses this contiguous array using a hashtable The input values are stored in order of ascending hash value, with conflicts stored adjacently. For each hash value, the position of the first value with that hash value is stored. The lookup then searches for a given value between the first and the last stored value with the same hash value.

Implementation for sparsely stored 6-index objects

We assume order [beta_1, alpha_1, beta_2, alpha_2, …]

A class for looping over tokens parsed from semantic lines.

The precomputed PropVec uniform excitation generator

This class efficiently returns unoccupied orbitals from an ilut-bitmask.