Edit on GitHub

iQIST recipes

Using the iQIST software package is quite easy. Next we will show you the standard workflow for using the iQIST software package.

Choose suitable component

At first, there are several CT-HYB quantum impurity solvers in the package. Their features and efficiency are somewhat different. Thus, it is the user's responsibility to choose suitable CT-HYB components to deal with the impurity problem at hand.

See also:

Design the programs and scripts

Second, the iQIST software package is in essence a computational engine, so users have to write scripts or programs to execute the selected CT-HYB impurity solver directly or to call it using the application programming interface. For example, if the users want to conduct CT-HYB/DMFT calculations, in principle they must implement the DMFT self-consistent equation by themselves.

There is a bonus. When the users want to study the Hubbard model on Bethe/cubic lattice using the single-site dynamical mean-field theory, or solve the Anderson impurity models in one-shot mode, it is possible to execute the quantum impurity solvers directly without any additional scripts or programs.

See also:

Prepare the input files

Third, an important task is to prepare proper input data for the selected CT-HYB impurity solver. The optional inputs for the CT-HYB impurity solver are the hybridization function [$\Delta(i\omega_n)$], impurity energy level ($E_{\alpha\beta}$), interaction parameters ($U$, $J$, $\lambda$, and $\mu$), etc. If users do not provide them to the impurity solver, it will use the default settings automatically. Specifically, if the Coulomb interaction matrix is general or the spin-orbital coupling is considered, users should use the JASMINE component to solve the local atomic Hamiltonian problem at first to generate the necessary eigenvalues and eigenvectors.

See also:

Let's go.

Fourth, execute the CT-HYB impurity solver directly or via some external scripts/programs.

See also:

Post-processing

Finally, when the calculations are finished, users can use the tools contained in the HIBISCUS component to post-process the output data, such as the imaginary-time Green's function $G(\tau)$, Matsubara self-energy function $\Sigma(i\omega_n)$, and other physical observables.

See also:

Good luck to you.