added support for macbook ARM64 processors.

added python 3.10 support.

added note to warning to notify users that hermiticity/symmetry checks are not exhaustive.

added small fixes to Optical Bloch equations: Lindblad dynamics using the matvec function.

fixed bug with automatic symmetry checks for *_basis_1d.

fixed bug with data types in quantum_operator for systems of N>31 sites.

fixed bug with using Nup argument of *_spin_basis_*.

fixed bug with return_rdm optional argument of tensor_basis.

fixed bug with Floquet tool returning the transposed Floquet evolution operator for complex-valued Floquet Hamiltonians.

updated tests removing deprecation warnings and errors.

new Example Liouville-von Neumann Equation using the MKL-enhanced Sparse Matrix Product [courtesy of J. Verlage] shows how to solve the Liouville-von Neumann equation using sparse matrices.

new Example Kitaev’s honeycomb Hamiltonian in vortex sectors given by the plaquette operators shows how to define symmetries using the user_basis that cannot be defined using the basis_general classes.

adding python 3.9 support.

fixed a bug with non-contiguous subsystems in basis.partial_trace() and basis.ent_entropy() for the fermionic basis clases *_fermion_basis_*.

fixed a bug with defining mixed particle sectors in *_fermion_basis_general.

fixed some typos in the documentation.

new issues templates on Github.

added optional argument noncommuting_bits to user_basis to specify the bits that represent a fermion degree of freedom.

adding python 3.8 support

adding quspin application to test quspin on hpc clusters.

added 2d array/batch support to tools.evolution.expm_multiply_parallel.

fixed errors caused by new releases of some of the dependencies.

fixed bugs with:

• particle-hole symmetry in *_fermion_basis_general;

• memory leakage in tools.lanczos;

• the reduce_output argument of basis_general.Op_bra_ket().

deprecated function tools.misc.csr_matvec.

added Majorana fermion operator strings to the *_fermion_basis_general.

added int_to_state and state_to_int functions to spinful_fermion_basis_* classes.

added properties shape and ndim to classes in the operator module.

new examples:

• Majorana fermion operators, cf. Gell-Mann Operators for Spin-1 systems;

• Gell-Mann operators for spin-1 systems, cf. Majorana Fermions: Spinless Fermi-Hubbard Model;

• Majorana SYK model, cf. Majorana Fermions: SYK Model.

• Calculation of spectral functions using symmetries, cf. Direct Calculation of spectral functions using symmetries.

• Tutorial on using QuSpin basis objects, cf. Basics of QuSpin basis objects.

discontinued support for python 2.7. Installing QuSpin for py27 will by defult result in version 0.3.3.

deprecated function basis.get_vec(): use basis.project_from() instead.

revised user_basis tutorial for spinless fermions and introduced function _count_particles_32().

added optional arguments svd_solver, svd_kwargs to basis.ent_entropy(); allows to use some scipy svd solvers, which are typically more stable.

expm_multiply_parallel now supports the option to give the operator an explicit dtype, see example Efficient time evolution using the omp-parallelized expm_multiply_parallel.

fixed bugs:

• computing the entanglement entropy when using the spinful_fermion_basis_general.

• constructing operators for higher-spin operators (S>1/2).

new *_basis_general functions – basis.project_from() and its inverse basis.project_to() – to transform states between a symmetry-reduced basis and the full basis.

new tools.Lanczos module for Lanczos type calculations, see examples Lanczos module: time-evolution and ground state search, Lanczos module: finite-temperature calculations.

new function method Op_shift_sector of the *basis_general_ classes allows to apply operators, which do not preserve the symmetry sector, to quantum states in the reduced basis. Useful for computing correlation functions. See example Autocorrelation functions using symmetries.

new required support package for QuSpin: numexpr.

introducing improvements to Example Monte Carlo Sampling Expectation Values of Obsevables to perform Monte Carlo sampling in the symmetry-reduced Hilbert space.

new examples:

• Example Fermi-Hubbard model without doubly-occupied sites to showcase double_occupancy option of the spinful_fermion_basis_*.

• Examples Quantum scars Hamiltonian: spin-1/2 system with Hilbert space constraint, Spin-1/2 system with sublattice particle consevation, Applying symmetries to reduce user-imported bases using QuSpin demonstrate the usage of user_basis.

• Example Optical Bloch equations: Lindblad dynamics using the matvec function shows how to use QuSpin for Lindblad dynamics and demonstrates the use of the omp-parallelzied matvec function for speedup.

• Example Hexagonal Lattice: Fermi-Hubbard model [courtesy of A. Buyskikh] shows how to construct Hamiltinians on a hexagonal lattice.

improved functionality of the tools.evolution.evolve() function.

fixed import issue with scipy’s comb function.

fixed a number of small bugs.

adding *_basis_general.get_amp() function method which effectively provides a partial get_vec() function but does not require the basis to be constructed ahead of time.

adding optional argument double_occupancy to the spinful_fermion_basis_* classes to control whether doubly occupied sites should be part of the basis or not.

adding the user_basis class which enables the user to build in Hilbert-space constraints, and exposes the inner workings of QuSpin’s core function to give the user almost complete control (see A tutorial on QuSpin’s user_basis).

adding tools.misc.matvec() and tools.misc.get_matvec() functions with omp-parallelized implementation which outperforms scipy and numpy in computing matrix-vector peroducts.

adding optional arguments to the dot() and rdot() functions of the operators module.

improved performance for matrix vector product in _oputils and expm_multiply_parallel. Leads to significant speedup in the hamiltonian and quantum_operator classes (e.g. in the hamiltonian.evolve() function) and the tools.evolution.expm_multiply_parallel() function.

support for python 3.7.

discontinued support for python 3.5 on all platforms and python 2.7 on windows. QuSpin for these versions will remain available to download up to and including QuSpin 0.3.0, but they are no longer being maintained.

matplotlib is no longer a required package to install quspin. It is still required to run the examples, though.

parallelization: New parallel features added or improved + OpenMP support for osx. Requires a different build of QuSpin (see also Parallel Computing Support).

new OpenMP features in operators module (see Parallel Computing Support and example script Parallel Computing in QuSpin).

improved OpenMP features in the *_general_basis classes.

new example scripts: (i) use of some new *_basis_general methods, (ii) use of OpenMP and QuSpin’s parallel features.

faster implementation of spin-1/2 and hard-core bosons in the general basis classes.

more memory efficient versions of matrix-vector/matrix products implemented for both hamiltonian and quantum_operator classes. Allows using OpenMP in the hamiltonian.evolve() function method.

refactored code for *_general_basis classes.

large integer support for *_general_basis classes allows to build lattices with more than 64 sites.

new argument make_basis for *_basis_general classes allows to use some of the basis functionality without constructing the basis.

new *_basis_general class methods: Op_bra_ket(), representative(), normalization(), inplace_Op().

support for Quantum Computing definition of “+”, “-“ Pauli matrices: see pauli argument of the spin_basis_* classes.

adding argument p_con to *_basis_general.get_vec() and *_basis_general.get_proj() functions.

adding functions basis.int_to_state() and basis.state_to_int() to convert between spin and integer representation of the states.

new basis.states attribute to show the list of basis states in their integer representation.

new methods of the *_basis_general classes for bitwise operations on basis states stored in integer representation.

both hamiltonian and quantum_operator classes support a new out argument for dot and rdot which allows the user to specify an output array for the result.

both hamiltonian and quantum_operator classes support a new overwrite_out argument which allows the user to toggle between overwriting the data within out or adding the result to out inplace without allocating extra data.