# DMFT: List of parameters

From ALPS

**Physical parameters**

- U: the Hubbard interaction U
- BETA: the inverse temperature
- MU: the chemical potential
- H: the magnetic field in the quantization axis (conventionally ) direction (BUT: the solvers do ignore the variable!)
- SITES: number of impurity sites (for DMFT: 1)
- FLAVORS: number of flavors/orbitals of the impurity (commonly 2: spin up/down)
- t: in case of Bethe lattice it does provide the hopping (the bandwidth is then , the half-bandwidth is ); if the option TWODBS is switched on then it does set the nearest-neighbor hopping on the square or hexagonal lattice
- t0, t1, ...: (available currently only for selfconsistency loop in imaginary time) sets the hopping for the Bethe lattice in multiband case (flavors and share the same parameter t
- J: coupling for the multiband problems
- U': (by default
- tprime: applies only if the option TWODBS is switched on and only for the square lattice, then it does set the next-nearest-neighbor hopping
- TWODBS: (by default sets the square lattice) you may choose either
`square`or`hexagonal`lattice

**Parameters for the self-consistency loop**

- OMEGA_LOOP: set it 1 unless you want to work with semicircular density of states (corresponding to the Bethe lattice in infinitely many dimensions)
- ANTIFERROMAGNET: if 1 then the antiferromagnetic self-consistency loop will be employed (formula 97 in review '96 of A.Georges et al)
- SYMMETRIZATION: if 1 then paramagnetic solution is enforced (in versions before 2.1: there has been a misspelling SYMMATRIZATION at several places and a usage of both, SYMMETRIZATION and SYMMATRIZATION, set to the same value was required)
- MAX_IT: maximum number of iteration in self-consistency loop (usually 10-20 will be enough)
- CONVERGED: criterium for stopping the self-consistency loop before reaching MAX_IT - if the maximum change in Green's function in Matsubara representation is less than CONVERGED, the loop will stop
- TOLERANCE: (only for hirschfyesim) as above
- RELAX_RATE: (by default 1; currently implemented only for selfconsistency loop with OMEGA_LOOP switched on) the new Green's function are in general computed as , which may help if oscillations occur

**General parameters**

- GENERAL_FOURIER_TRANFORMER: set it on if you have OMEGA_LOOP and other than the Bethe lattice
- EPS_i: (i=0,1,...,FLAVORS-1) potential shift for the flavor (necessary for GENERAL_FOURIER_TRANSFORMER)
- EPSSQ_i: (i=0,1,...,FLAVORS-1) the second moment of the bandstructure for the flavor (necessary for GENERAL_FOURIER_TRANSFORMER)
- DOSFILE: sets the name for the file containing the density of states (expected 2 columns with energy value and corresponding density of states at that energy; equidistant energies required; odd number of rows required due to Simpson integration)
- TWODBS: switches on the Hilbert transformation for 2-dimensional systems, currently supported square lattice (with nearest and next-nearest neighbor hoppings) and hexagonal lattice (with nearest neighbor hoppings) [Note: a different 2-dimensional lattice may be easily added]
- L: optional parameter available in case of TWODBS is on; defines the half of the linear discretization in the integration in the self-consistency (default: 200)
- SOLVER: specifies the impurity solver ("Hybridization" or "Interaction Expansion"; the solver "Hirsch-Fye" does suffer from discretization errors and is thus not recommended)

**Parameters for the initial/final Weiss field**

- H_INIT: magnetic field in the quantization axis (conventionally ) direction, which is used in computation of the non-interacting initial G0 (if it is not loaded)
- G0OMEGA_INPUT: name for the text file specifying the Weiss field in Matsubara frequencies (expected columns, and total NMATSUBARA rows) (use only with OMEGA_LOOP)
- G0TAU_INPUT: name for the text file specifying the Weiss field in imaginary time representation (expected columns, and total rows) (only with OMEGA_LOOP switched off)
- GOMEGA_input: specifies the name for the text file where the initial G0 in Matsubara representation will be written (by default it is not written, as it is identical with
`G0_omega_1`) - G0TAU_input: name for the text file for the output of the initial G0 in imaginary time (by default it is not written, as it is identical with
`G0_tau_1`) - G0OMEGA_output: name for the output file containing the final Weiss field in Matsubara frequencies (by default
`G0omega_output`)(with OMEGA_LOOP) - G0TAU_output: name for the output file containing the final Weiss field in Matsubara frequencies (by default
`G0tau_output`) (with OMEGA_LOOP off) - INSULATING: if you have specified this option, then the initial G0 will be set up in the insulating limit

**Parameters setting the precision of representation of the Green's function and the Weiss field**

- NMATSUBARA: number of Matsubara frequencies used to represent the Green's function and the Weiss field (usually equals N)
- N: number of bins for the Green's function and the Weiss field in imaginary time (represented in total by N+1 values) (recommended: roughly 1000 for the continuous-time solvers)

**Hybridization1 expansion impurity solver parameters**

- MAX_TIME: sets the maximum time given in seconds spent on the impurity problem solving (basically this sets the duration of a single iteration)
- SWEEPS: number of desired sweeps performed during the calculation (recommendation: set it very high, e.g. and the solver will stop on the time limit given by MAX_TIME)
- THERMALIZATION: number of sweeps before the Monte Carlo measurements in order to reach configuration close to equilibrium (of the order of 1000)
- EPSSQAV: the second moment of the bandstructure (necessary if you have specified your own DOSFILE)
- N_ORDER: setting histogram size (if the hybridization order is larger then it will be not stored in the histogram) (value of the order of 100 might be reasonable)
- N_MEAS: number of Monte Carlo steps between measurements (of the order of 10000)
- N_SHIFT: number of shifts of segments in a single Monte Carlo step (apparently unused, so 0)
- MEASURE_FOURPOINT: if switched on then the four-point correlators are being measured
- N4point: (only used if MEASURE_FOURPOINT is on)
*description missing so far* - CHECKPOINT: filename prefix for checkpointing files and for the final h5 and xml output

**Interaction expansion1 impurity solver parameters**

- MAX_TIME: sets the maximum time given in seconds spent on the impurity problem solving
- SWEEPS: number of desired sweeps performed during the calculation (recommendation: set it very high, e.g. and the solver will stop on the time limit given by MAX_TIME)
- THERMALIZATION: number of sweeps before the Monte Carlo measurements in order to reach configuration close to equilibrium (of the order of 1000)
- SWEEP_MULTIPLICATOR: (default: 1)
- NRUNS: (default: 1)
- ALPHA:
- RECALC_PERIOD: (default: 5000)
- MEASUREMENT_PERIOD: (default: 200)
- CONVERGENCE_CHECK_PERIOD: (default provided)
- ALMOSTZERO: (default: )
- NSELF: (default: )
- NMATSUBARA_MEASUREMENTS: (default: NMATSUBARA)
- HISTOGRAM_MEASUREMENT: (default: false)
- GET_COMPACTED_MEASUREMENTS:
- ATOMIC:
- TAU_DISCRETIZATION_FOR_EXP:
- CHECKPOINT: filename prefix for the checkpointing files and for the final h5 and xml output

**Additional parameters**

- SEED: random seed for the pseudorandom generator
- RNG: pseudorandom generator used (default is "mt19937"), might be switched to "lagged_fibonacci607"