Skip to content

rlx input variables

This document lists and provides the description of the name (keywords) of the rlx input variables to be used in the input file for the abinit executable.

adpimd

Mnemonics: ADiabatic Path-Integral Molecular Dynamics
Mentioned in topic(s): topic_PIMD
Variable type: integer
Dimensions: scalar
Default value: 0
Only relevant if: imgmov == 9 or imgmov == 13
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Controls whether adiabatic Path-Integral Molecular Dynamics is performed or not. The corresponding adiabaticity parameter is given by adpimd_gamma.

If equal to 0, no adiabatic Path-Integral Molecular Dynamics (standard PIMD) is performed. If equal to 1, adiabatic Path-Integral Molecular Dynamics is activated. Only relevant with pitransform = 1 (normal mode transformation). In that case,

  • the mass associated with the zero-frequency mode is the true mass amu,
  • the mass associated to the other higher frequency modes of the polymer chains is equal to the normal mode mass divided by adpimd_gamma (adiabaticity parameter),
  • the equation of motion on the zero-frequency mode is not thermostated.

NOT YET USABLE

adpimd_gamma

Mnemonics: ADiabatic Path-Integral Molecular Dynamics: GAMMA factor
Mentioned in topic(s): topic_PIMD
Variable type: real
Dimensions: scalar
Default value: 1
Only relevant if: adpimd == 1 and imgmov in [9,13]
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Adiabaticity parameter to be used in adiabatic Path-Integral Molecular Dynamics. NOT YET USABLE

amu

Mnemonics: Atomic Mass Units
Characteristics: EVOLVING
Mentioned in topic(s): topic_PIMD, topic_Phonons, topic_AtomTypes, topic_Artificial
Variable type: real
Dimensions: (ntypat)
Default value: None
Comment: Defaults provided by a database of atomic masses.
Added in version: before_v9

Test list (click to open). Moderately used, [82/1245] in all abinit tests, [1/159] in abinit tutorials

Gives the masses in atomic mass units for each kind of atom in the input cell. These masses are used in performing molecular dynamical atomic motion if ionmov = 1, 6, 7 or 8. They are also used in phonon calculations during the diagonalization of the dynamical matrix. Note that one may set all masses to 1 for certain cases in which merely structural relaxation is desired and not actual molecular dynamics.

Using the recommended values of [Martin1987], 1 atomic mass unit = 1.6605402e-27 kg. In this unit the mass of Carbon 12 is exactly 12.

A database of atomic masses is provided which provides the default values. Note that the default database uses mixed isotope masses (for Carbon the natural occurrence of Carbon 13 is taken into account). The values are those recommended by the commission on Atomic Weights and Isotopic Abundances, Inorganic Chemistry Division, IUPAC [Martin1987]. For Tc, Pm, Po to Ac, Pa and beyond U, none of the isotopes have a half-life greater than 3.0d10 years, and the values provided in the database do not come from that source.

For alchemical pseudoatoms, the masses of the constituents atoms are mixed, according to the alchemical mixing coefficients mixalch

In most cases, the use of amu will be as a static (non-evolving) variable. However, the possibility to have different values of amu for different images has been coded. A population of cells with different atomic characteristics can thus be considered, and can be made to evolve, e.g. with a genetic algorithm (not coded in v7.0.0 though).

bmass

Mnemonics: Barostat MASS
Mentioned in topic(s): topic_MolecularDynamics
Variable type: real
Dimensions: scalar
Default value: 10
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

bmass is the mass of the barostat when ionmov = 13 (constant pressure molecular dynamics)

chkdilatmx

Mnemonics: CHecK DILATMX
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [8/1245] in all abinit tests, [0/159] in abinit tutorials

If 0, the code will not stop execution if the threshold of dilatmx is exceeded, it will simply issue a warning. There will be no rescaling. If 1, after tentative rescaling as described in dilatmx, the code will stop execution. Also, the use of chkdilatmx = 0 allows one to set dilatmx to a larger value than 1.15, otherwise forbidden as being a waste of CPU and memory.

So, when using chkdilatmx =0, the relaxed lattice parameters might not be accurate, but will simply better than the starting ones.

chkdilatmx =0 is useful when the starting geometry is likely very inaccurate. However, if the user is in search of an accurate geometry estimation, then a first determination of the (better but inaccurate) geometry with chkdilatmx =0 should be followed by a more accurate second run from the better geometry with chkdilatmx =1 and dilatmx slightly larger than 1, (possibly 1.05).

cineb_start

Mnemonics: Climbing-Image Nudged Elastic Band: STARTing iteration
Mentioned in topic(s): topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 7
Only relevant if: imgmov == 5 and neb_algo == 2
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the index of the first CI-NEB iteration. The CI-NEB method constitutes a small modification to the NEB method allowing a rigorous convergence to the saddle point. As the image with the highest energy has to be identified, the calculation begins with several iterations of the standard NEB algorithm. The effective CI-NEB begins at the cineb_start iteration. See [Henkelman2000a] for additional details of this method.

delayperm

Mnemonics: DELAY between trials to PERMUTE atoms
Mentioned in topic(s): topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Delay (number of time steps) between trials to permute two atoms, in view of accelerated search of minima. Still in development.

See the routine moldyn.F90 and signperm for additional information.

When delayperm is zero, there are no permutation trials.

diismemory

Mnemonics: Direct Inversion in the Iterative Subspace MEMORY
Mentioned in topic(s): topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 8
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the maximum number of “time” steps for which the forces and stresses are stored, and taken into account in the DIIS algorithm (ionmov = 20) to find zero-force and stress configurations.

dilatmx

Mnemonics: lattice DILATation: MaXimal value
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 1.0
Added in version: before_v9

Test list (click to open). Moderately used, [71/1245] in all abinit tests, [15/159] in abinit tutorials

dilatmx is an auxiliary variable used to book additional memory (see detailed description later) for possible on-the-flight enlargement of the plane wave basis set, due to cell volume increase during geometry optimization by ABINIT. Useful only when doing cell optimization, e.g. optcell/=0, usually with ionmov == 2 or 22. Supposing that the starting (estimated) lattice parameters are already rather accurate (or likely to be too large), then the recommended value of dilatmx is 1.05. When you have no idea of evolution of the lattice parameters, and suspect that a large increase during geometry optimization is possible, while you need an accurate estimation of the geometry, then make a first run with chkdilatmx=0, producing an inaccurate, but much better estimation, followed by a second run using the newly estimated geometry, with chkdilatmx=1 (the default) and dilatmx set to 1.05. If you are not in search of an accurate estimation of the lattice parameters anyhow, then run with chkdilatmx=0 only once.

In the default mode (chkdilatmx = 1), when the dilatmx threshold is exceeded, ABINIT will rescale uniformly the tentative new primitive vectors to a value that leads at most to 90% of the maximal allowed dilatmx deviation from 1. It will do this three times (to prevent the geometry optimization algorithms to have taken a too large trial step), but afterwards will stop and exit.

Setting chkdilatmx == 0 allows one to book a larger planewave basis (if dilatmx is set to be bigger than 1), but will not rescale the tentative new primitive vectors nor lead to an exit when the dilatmx threshold is exceeded. The obtained optimized primitive vectors will not be exactly the ones corresponding to the planewave basis set determined using ecut at the latter primitive vectors. Still, as an intermediate step in a geometry search this might be sufficiently accurate. In such case, dilatmx might even be let at its default value 1.0.

Detailed explanation: The memory space for the planewave basis set is defined by multiplying ecut by dilatmx squared (the result is an “effective ecut”, called internally “ecut_eff”). Other uses of ecut are not modified when dilatmx > 1.0. Still, operations (like scalar products) are done by taking into account these fake (non-used) planewaves, even if their coefficients are set to zero, thus slowing down the ABINIT execution. Using dilatmx <1.0 is equivalent to changing ecut in all its uses. This is allowed, although its meaning is no longer related to a maximal expected scaling.

Setting dilatmx to a large value leads to waste of CPU time and memory. By default, ABINIT will not accept that you define dilatmx bigger than 1.15. This behaviour will be overcome by using chkdilatmx == 0. Supposing you think that the optimized acell values might be 5% larger than your input values, use simply dilatmx 1.05. This will lead to an increase of the number of planewaves by a factor (1.05)^3, which is about 1.158, and a corresponding increase in CPU time and memory. It is possible to use dilatmx when optcell =0, but a value larger than 1.0 will be a waste.

dtion

Mnemonics: Delta Time for IONs
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics, topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 100
Added in version: before_v9

Test list (click to open). Moderately used, [24/1245] in all abinit tests, [1/159] in abinit tutorials

Used for controlling ion time steps. If ionmov is set to 1, 6, 7 and 15, then molecular dynamics is used to update atomic positions in response to forces. The parameter dtion is a time step in atomic units of time. (One atomic time unit is 2.418884e-17 seconds, which is the value of Planck’s constant in hartree*sec.) In this case the atomic masses, in amu (given in array ” amu “), are used in Newton’s equation and the viscosity (for ionmov =1) and number of time steps are provided to the code using input variables “vis” and “ntime”. The code actually converts from masses in amu to masses in atomic units (in units of electron masses) but the user enters masses in amu. (The conversion from amu to atomic units (electron masses) is 1822.88851 electron masses/amu.)

A typical good value for dtion is about 100. The user must try several values for dtion in order to establish the stable and efficient choice for the accompanying amu, atom types and positions, and vis (viscosity). For quenched dynamics (ionmov = 7), a larger time step might be taken, for example 200. No meaning for RF calculations. It is also used in geometric relaxation calculation with the FIRE algorithm (ionmov=15), where the time is virtual. A small dtion should be set, for example 0.03.

dynimage

Mnemonics: list of DYNamic IMAGEs
Mentioned in topic(s): topic_PIMD, topic_TransPath
Variable type: integer
Dimensions: (nimage)
Default value: 1
*Comment: if imgmov in [2,5] (String Method, NEB), dynimage(1)=0 and dynimage(nimage)=0.
Added in version: before_v9

Test list (click to open). Moderately used, [13/1245] in all abinit tests, [2/159] in abinit tutorials

This input variable is relevant when sets of images are activated (see imgmov). Not all images might be required to evolve from one time step to the other. Indeed, in the String Method or the Nudged Elastic Band, one might impose that the extremal configurations of the string are fixed. In case dynimage (iimage)=0, the image with index “iimage” will be consider as fixed. Thus, there is no need to compute forces and stresses for this image at each time step. The purpose of defining extremal images is to make the input/output easier.

In order to save CPU time, the computation of properties of static images ( dynimage (iimage)=0) can be avoided: see istatimg keyword.

ecutsm

Mnemonics: Energy CUToff SMearing
Characteristics: ENERGY
Mentioned in topic(s): topic_Planewaves
Variable type: real
Dimensions: scalar
Default value: 0.0
Added in version: before_v9

Test list (click to open). Moderately used, [148/1245] in all abinit tests, [24/159] in abinit tutorials

This input variable is important when performing relaxation of unit cell size and shape (non-zero optcell). Using a non-zero ecutsm , the total energy curves as a function of ecut, or acell, can be smoothed, keeping consistency with the stress (and automatically including the Pulay stress).

The recommended value is 0.5 Ha in such a case (non-zero optcell).

Actually, when optcell/=0, ABINIT requires ecutsm to be larger than zero. If you want to optimize cell shape and size without smoothing the total energy curve (a dangerous thing to do), use a very small ecutsm , on the order of one microHartree.

Technical information: See Appendix B of [Laflamme2016]. ecutsm allows one to define an effective kinetic energy for plane waves, close to, but lower than, the maximal kinetic energy ecut. For kinetic energies less than ecut- ecutsm , nothing is modified, while between ecut- ecutsm and ecut, the kinetic energy is multiplied by 1.0 / ( x^2(3+x-6x^2+3x^3 )), where x = (ecut - kinetic_energy)/ ecutsm . Note that x^2(3+x-6x^2+3x^3) is 0 at x=0, with vanishing derivative, and that at x=1, it is 1, with also vanishing derivative. If ecutsm is zero, the unmodified kinetic energy is used. ecutsm can be specified in Ha (the default), Ry, eV or Kelvin, since ecutsm has the ENERGY characteristics. (1 Ha = 27.2113845 eV). A few test for Silicon (diamond structure, 2 k-points) have shown 0.5 Ha to be largely enough for ecut between 2Ha and 6Ha, to get smooth curves. It is likely that this value is OK as soon as ecut is larger than 4Ha.

friction

Mnemonics: internal FRICTION coefficient
Mentioned in topic(s): topic_MolecularDynamics
Variable type: real
Dimensions: scalar
Default value: 0.001
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the internal friction coefficient (atomic units) for Langevin dynamics (when ionmov = 9): fixed temperature simulations with random forces.

The equation of motion is: M_I \frac{d^2 R_I}{dt^2}= F_I - friction *M_I \frac{d R_I}{dt} - F_{random,I}, where F_{random,I} is a Gaussian random force with average zero and variance friction *2M_IkT. The atomic unit of friction is Hartree*Electronic mass*(atomic unit of Time)/Bohr^2. See [Chelikowsky2000] for additional information.

fxcartfactor

Mnemonics: Forces to (X) CARTesian coordinates FACTOR
Mentioned in topic(s): topic_TransPath, topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 1 (Bohr^2)/Hartree
Added in version: before_v9

Test list (click to open). Rarely used, [9/1245] in all abinit tests, [0/159] in abinit tutorials

The forces multiplied by fxcartfactor will be treated like difference in cartesian coordinates in the process of optimization. This is a simple preconditioner. TO BE UPDATED See (ionmov = 2 or 22, non-zero optcell). For example, the stopping criterion defined by tolmxf relates to these scaled stresses.

ga_algor

Mnemonics: Genetic Algorithm - ALGOrithm selection
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

Choosing method to make the structure selection. Only the enthalpy is used now but we plan to include, energy, electronic band gap and alchemical potentials. Right now only value of 1 (enthalpy) works.

ga_fitness

Mnemonics: Genetic Algorithm FITNESS function selection
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Different methodologies to perform the roulette-wheel selection of parents. Even though, the objective function is the crystalline enthalpy (H_i), the weight of the population elements to be chosen from in a roulette-wheel selection can be given through different functions. We consider the following cases.

  1. F = H_i / Sum H_i
  2. F = exp(-(H_i-H_min)) / Sum exp(-(H_i-H_min))
  3. F = (1/n_i) / Sum (1/n_i). Where n_i is the position in the ordered list of enthalpies

ga_n_rules

Mnemonics: Genetic Algorithm Number of RULES
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Different genetic rules have been implemented and the user has the change to choose between any of them. Right now we have 4 rules. See ga_rules

ga_opt_percent

Mnemonics: Genetic Algorithm OPTimal PERCENT
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 0.2
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Percentage of the population that according to the fitness function passes to the following iteration.

ga_rules

Mnemonics: Genetic Algorithm RULES
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Different genetic rules have been implemented and the user has the change to choose between any of them. The chosen number of rules have been defined in ga_n_rules

Implemented rules are 1) crossover. Two parents are randomly chosen and two springs are mixed from the two by (a) choosing randomly (through Fitness function) two parents and then randomly rotating and shifting the coordinates withing that particular cell. (b) Slice every one of the unit cell of the parents along a random direction and creating the spring offs from the pieces of the two parents. 2) Vector flip mutation. From the coordinates from a given parent, a piece of it is inverted. 3) Random strain. A random anisotropic deformation is given to the unit cell. 4) Coordinates mutation of ¼ of the whole coordinates.

getcell

Mnemonics: GET CELL parameters from…
Mentioned in topic(s): topic_multidtset
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [5/1245] in all abinit tests, [1/159] in abinit tutorials

This variable is typically used to chain the calculations, in the multi- dataset mode (ndtset > 0). It describes from which dataset the output acell and rprim are to be taken (implicitly also scalecart), as input of the present dataset. The cell parameters are EVOLVING variables, for which such a chain of calculations is useful.

  • If getcell =0, no previously computed values are used.
  • If getcell >0, the value must be the index of the dataset from which the cell data is to be used as input data. It must be the index of a dataset already computed in the SAME run.
  • If getcell =-1, the output data of the previous dataset must be taken, which is a frequently occurring case. However, if the first dataset is treated, -1 is equivalent to 0, since no dataset has been computed in the same run.
  • If getcell =another negative number, it indicates the number of datasets to go backward to find the needed data (once again, going back beyond the first dataset is equivalent to using a zero get variable).

getvel

Mnemonics: GET VEL from…
Mentioned in topic(s): topic_multidtset
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

This variable is typically used to chain the calculations, in the multi- dataset mode (ndtset > 0) since they describe from which dataset the corresponding output variables are to be taken, as input of the present dataset. The velocities are EVOLVING variables, for which such a chain of calculation is useful.

  • If getvel == 0, no such use of previously computed vel is done.

  • If getvel is positive, its value gives the index of the dataset from which the data are to be used as input data. It must be the index of a dataset already computed in the SAME run.

  • If getvel is -1, the output data of the previous dataset must be taken, which is a frequently occurring case.

  • If getvel is a negative number, it indicates the number of datasets to go backward to find the needed data. Going back beyond the first dataset is equivalent to using zero for the get variable.

Note: The use of getvel with getxred is allowed, despite the different coordinate system.

getxcart

Mnemonics: GET XCART from…
Mentioned in topic(s): topic_multidtset
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [8/1245] in all abinit tests, [0/159] in abinit tutorials

This variable is typically used to chain the calculations, in the multi- dataset mode (ndtset > 0) since it describes from which dataset the corresponding output variables are to be taken, as input of the present dataset. The atomic positions are EVOLVING variables, for which such a chain of calculation is useful.

Note that the use of getxcart and getxred differs when acell and rprim are different from one dataset to the other.

  • If 0, no previously computed values are used.

  • If >0, the integer must correspond to the index of the dataset from which the data are to be used as input data. It must be the index of a dataset already computed in the SAME run.

  • If equal to -1, the output data of the previous dataset must be taken, which is a frequently occurring case. However, if the first dataset is treated, -1 is equivalent to 0, since no dataset has yet been computed in the same run.

  • If another negative number, it indicates the number of datasets to go backward to find the needed data. Going back beyond the first dataset is equivalent to using a null get variable.

Note: getxred and getxcart cannot be simultaneously non-zero for the same dataset. On the other hand the use of getvel with getxred is allowed, despite the different coordinate system.

getxred

Mnemonics: GET XRED from…
Mentioned in topic(s): topic_multidtset
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [12/1245] in all abinit tests, [1/159] in abinit tutorials

This variable is typically used to chain the calculations, in the multi- dataset mode (ndtset > 0) since it describes from which dataset the corresponding output variables are to be taken, as input of the present dataset. The atomic positions and velocities are EVOLVING variables, for which such a chain of calculation is useful.

Note that the use of getxcart and getxred differs when acell and rprim are different from one dataset to the other.

  • If 0, no previously computed values are used.

  • If >0, the integer must correspond to the index of the dataset from which the data are to be used as input data. It must be the index of a dataset already computed in the SAME run.

  • If equal to -1, the output data of the previous dataset must be taken, which is a frequently occurring case. However, if the first dataset is treated, -1 is equivalent to 0, since no dataset has yet been computed in the same run.

  • If another negative number, it indicates the number of datasets to go backward to find the needed data. Going back beyond the first dataset is equivalent to using a null get variable.

Note: getxred and getxcart cannot be simultaneously non-zero for the same dataset. On the other hand the use of getvel with getxred is allowed, despite the different coordinate system.

goprecon

Mnemonics: Geometry Optimization PRECONditioner equations
Mentioned in topic(s): topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Set the kind of preconditioner to be used for Geometry Optimization (Note: Under development now (2011.05.20))

  • goprecon = 0: No preconditioner
  • goprecon = [1-9]: Linear preconditioner
  • goprecon = [11-19]: Non-linear preconditioner

goprecprm

Mnemonics: Geometry Optimization PREconditioner PaRaMeters equations
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: (3)
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Set the parameters use by the preconditioner to be used for Geometry Optimization (Note: Under development now (2011.06.06))

hmcsst

Mnemonics: Hybrid Monte Carlo Strain Step Trajectory
Mentioned in topic(s): topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 0
Only relevant if: ionmov == 25
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

Number of strain teps per MC trial trajectory, for the Hybrid Monte Carlo algorithm ionmov=25.

hmctt

Mnemonics: Hybrid Monte Carlo Trial Trajectory
Mentioned in topic(s): topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 0
Only relevant if: ionmov == 25
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

Number of steps per MC trial trajectory, for the Hybrid Monte Carlo algorithm ionmov=25.

iatcon

Mnemonics: Indices of AToms in CONstraint equations
Characteristics: NO_MULTI, INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (natcon,nconeq)
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the indices of the atoms appearing in each of the nconeq independent equations constraining the motion of atoms during structural optimization or molecular dynamics (see nconeq, natcon, and wtatcon). (Note: combined with wtatcon to give internal representation of the latter)

iatfix

Mnemonics: Indices of AToms that are FIXed
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (natfix)
Default value: None
Only relevant if: natfix > 0
Added in version: before_v9

Test list (click to open). Moderately used, [32/1245] in all abinit tests, [6/159] in abinit tutorials

Give the index (in the range 1 to natom ) of each atom which is to be held fixed for structural optimization or molecular dynamics. The variable iatfix lists those fixed in the three directions, while the variables iatfixx, iatfixy, and iatfixz, allow to fix some atoms along x, y or z directions, or a combination of these.

WARNING: The implementation is inconsistent !! For ionmov ==1, the fixing of directions was done in cartesian coordinates, while for the other values of ionmov, it was done in reduced coordinates. Sorry for this.

There is no harm in fixing one atom in the three directions using iatfix , then fixing it again in other directions by mentioning it in iatfixx, iatfixy or iatfixz. The internal representation of these input data is done by the mean of one variable iatfix (3,natom), defined for each direction and each atom, being 0 if the atom is not fixed along the direction, and 1 if the atom is fixed along the direction. When some atoms are fixed along 1 or 2 directions, the use of symmetries is restricted to symmetry operations whose (3x3) matrices symrel are diagonal. If the atom manipulator is used, iatfix will be related to the preprocessed set of atoms, generated by the atom manipulator. The user must thus foresee the effect of this atom manipulator (see objarf).

iatfixx

Mnemonics: Indices of AToms that are FIXed along the X direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (natfixx)
Default value: None
Only relevant if: natfixx > 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Give the index (in the range 1 to natom ) of each atom which is to be held fixed ALONG THE X direction for structural optimization or molecular dynamics. The variable iatfix lists those fixed in the three directions, while the variables iatfixx , iatfixy, and iatfixz, allow to fix some atoms along x, y or z directions, or a combination of these. See the variable iatfix for more information.

iatfixy

Mnemonics: Indices of AToms that are FIXed along the Y direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (natfixy)
Default value: None
Only relevant if: natfixy > 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Give the index (in the range 1 to natom ) of each atom which is to be held fixed ALONG THE Y direction for structural optimization or molecular dynamics. The variable iatfix lists those fixed in the three directions, while the variables iatfixx, iatfixy , and iatfixz, allow to fix some atoms along x, y or z directions, or a combination of these. See the variable iatfix for more information.

iatfixz

Mnemonics: Indices of AToms that are FIXed along the Z direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (natfixz)
Default value: None
Only relevant if: natfixz > 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Give the index (in the range 1 to natom ) of each atom which is to be held fixed ALONG THE Z direction for structural optimization or molecular dynamics. The variable iatfix lists those fixed in the three directions, while the variables iatfixx, iatfixy, and iatfixz , allow to fix some atoms along x, y or z directions, or a combination of these. See the variable iatfix for more information.

imgmov

Mnemonics: IMaGe MOVEs
Mentioned in topic(s): topic_CrossingBarriers, topic_PIMD, topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Moderately used, [19/1245] in all abinit tests, [5/159] in abinit tutorials

Control the collective changes of images (see nimage,npimage, dynimage, ntimimage, tolimg, istatimg, prtvolimg). Similar to ionmov in spirit, although here, a population of self-consistent calculations for possibly different (evolving) geometries is managed, while with ionmov, only self-consistent calculation for one (evolving) geometry is managed. In this respect the maximal number of time step for image propagation is ntimimage, corresponding to the input variable ntime of the single geometry case. Also, the stopping criterion is governed by tolimg, corresponding to the input variable toldfe of the single geometry case. The stopping condition is crude: the image propagation is stopped when the mean value (over dynamic images) of the absolute difference of total energy (previous and current time step) is less than tolimg.

Actually, there might be combinations of ionmov and imgmov in which the two mechanisms are at work. Usually, however, only one mechanism will be activated (so, usually, either ntimimage is bigger than one OR ntime is bigger than one). In order for the user to acquire a mental representation of the interplay between ionmov and imgmov , here is a F90 pseudo-code presenting the interplay between the different above-mentioned input variables, as well as with the parallelism (see input variable npimage).

    do itimimage=1,ntimimage
      do iimage=1,nimage
        (possibly, parallelisation over images)
        do itime=1,ntime
          Compute the forces and stresses for image(iimage)
          Examine whether the stopping criterion defined by tolmxf is fulfilled
          Predict the next geometry for image(iimage) using ionmov
        enddo
      enddo
      Examine whether the stopping criterion defined by tolimg is fulfilled
      Predict the next geometries for all images using imgmov
    enddo
  • = 0 → simply copy images from previous itimimage step.
  • = 1 → move images according to Steepest Descent following the (scaled) forces, the scaling factor being fxcartfactor.
  • = 2 → String Method for finding Minimal Energy Path (MEP) connecting two minima (see [Weinan2002]), or even two configurations that are not local minima; the algorithm variant can be selected with the string_algo keyword (Simplified String Method by default). The solver for the Ordinary Differential Equation (ODE) can be selected with mep_solver (steepest-descent by default). See also mep_mxstep keyword.
  • = 3 → (tentatively, not yet coded) Metadynamics.
  • = 4 → (tentatively, not yet coded) Genetic Algorithm.
  • = 5 → Nudged Elastic Band (NEB) for finding Minimal Energy Path (MEP) connecting two minima; the algorithm variant can be selected with the neb_algo keyword (NEB+improved tangent by default). The solver for the Ordinary Differential Equation (ODE) can be selected with mep_solver (steepest-descent by default). The spring constant connecting images along the path is defined by neb_spring. See also mep_mxstep keyword.
  • = 6 → Linear Combination of Constrained DFT Energies. The images can have different electronic structure (occ can differ), and their total energies are combined linearly using the factors in mixesimgf, giving the actual total energy of the ensemble of constrained DFT images. The geometry is the same for all images, forces and stresses are computed, and all usual algorithms for MD or geometry optimization are allowed, using ionmov (instead of imgmov , this is the exception to the rule) and related variables.
  • = 9 or 13 → Path-Integral Molecular Dynamics (see e.g. [Marx1996]). Will use 9 for Langevin thermostat (associated friction coefficient given by vis) and 13 for Nose-Hoover thermostat chains (associated input variables are the number of thermostats in the chains, nnos, and the masses of these thermostats qmass). nimage is the Trotter number (no use of dynimage); possible transformations of coordinates are defined by pitransform; Fictitious masses of the atoms (possibly different from the true masses given by amu) can be specified by pimass. At present, it is only possible to perform calculations in the (N,V,T) ensemble (optcell = 0).

No meaning for RF calculations.

imgwfstor

Mnemonics: IMaGe WaveFunction STORage
Mentioned in topic(s): topic_CrossingBarriers, topic_PIMD, topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 0
Only relevant if: extrapwf == 0 and ntimimage > 0
Added in version: before_v9

Test list (click to open). Rarely used, [6/1245] in all abinit tests, [3/159] in abinit tutorials

Govern the storage of wavefunctions at the level of the loop over images, see ntimimage. Possible values of imgwfstor are 0 or 1. If imgwfstor is 1, the wavefunctions for each image are stored in a big array of size nimage more than the storage needed for one set of wavefunctions. When the specific computation (optimization/SCF cycle etc) for this image is started, the past wavefunctions are used, to speed up the computation. If imgwfstor ==0, the wavefunctions are reinitialised, either at random or from the initial wavefunction file (so, without any modification to take into account the computations at the previous value of itimimage.

If nimage is large, the increase of memory need can be problematic, unless the wavefunctions are spread over many processors, which happens when paral_kgb == 1. For some algorithms, e.g. when some geometry optimization is performed, imgmov==2 or 5, the gain in speed of choosing imgwfstor =1 can be quite large, e.g. two to four. For algorithms of the molecular dynamics type, imgmov==9 or 13, the expected gain is smaller. Of course, with adequate memory resources, imgwfstor ==1 should always be preferred.

ionmov

Mnemonics: IONic MOVEs
Mentioned in topic(s): topic_MolecularDynamics, topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Moderately used, [145/1245] in all abinit tests, [24/159] in abinit tutorials

Choice of algorithm to control the displacements of ions, and possibly changes of cell shape and size (see optcell). No meaning for RF calculations.

  • 0 → Do not move ions (default behaviour)

  • 1 → Move atoms using molecular dynamics with optional viscous damping (friction linearly proportional to velocity). The viscous damping is controlled by the vis parameter. If undamped molecular dynamics is desired, set vis to 0. The implemented algorithm is the generalisation of the Numerov technique (6th order), but is NOT invariant upon time-reversal, so that the energy is not conserved. The value ionmov = 6 will usually be preferred, although the algorithm that is implemented is lower-order. The time step is governed by dtion. Purpose: Molecular dynamics (if vis = 0), Structural optimization (if vis >0) Cell optimization: No (Use optcell = 0 only) Related variables: Viscous parameter vis, time step dtion, index of atoms fixed iatfix

  • 2 → Conduct structural optimization using the Broyden-Fletcher-Goldfarb-Shanno minimization (BFGS). This is much more efficient for structural optimization than viscous damping, when there are less than about 10 degrees of freedom to optimize. Another version of the BFGS is available with ionmov == 22, and is apparently more robust and efficient than ionmov == 2. Purpose: Structural optimization Cell optimization: Yes (if optcell/=0)

  • 3 → Conduct structural optimization using the Broyden-Fletcher-Goldfarb-Shanno minimization (BFGS), modified to take into account the total energy as well as the gradients (as in usual BFGS). See [Schlegel1982]. Might be better than ionmov = 2 for few degrees of freedom (less than 3 or 4). Can be very unstable - use with caution! Purpose: Structural optimization Cell optimization: Yes (if optcell/=0)

  • 4 → Conjugate gradient algorithm for simultaneous optimization of potential and ionic degrees of freedom. It can be used with iscf = 2 and iscf =5 or 6 (WARNING: this is under development, and does not work very well in many cases). Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only)

  • 5 → Simple relaxation of ionic positions according to (converged) forces. Equivalent to ionmov = 1 with zero masses, albeit the relaxation coefficient is not vis, but iprcfc. Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only)

  • 6 → Molecular dynamics using the Verlet algorithm, see [Allen1987a] p 81]. The only related parameter is the time step (dtion). Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only) Related variables: time step dtion, index of atoms fixed iatfix

  • 7 → Quenched Molecular dynamics using the Verlet algorithm, and stopping each atom for which the scalar product of velocity and force is negative. The only related parameter is the time step (dtion). The goal is not to produce a realistic dynamics, but to go as fast as possible to the minimum. For this purpose, it is advised to set all the masses to the same value (for example, use the Carbon mass, i.e. set amu to 12 for all type of atoms). Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only) Related variables: time step dtion, index of atoms fixed iatfix

  • 8 → Molecular dynamics with Nose-Hoover thermostat, using the Verlet algorithm. Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only) Related variables: time step (dtion), Temperatures (mdtemp), and thermostat mass (noseinert).

  • 9 → Langevin molecular dynamics. Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only) Related variables: time step (dtion), temperatures (mdtemp) and friction coefficient (friction).

  • 10 → Delocalized internal coordinates with simple BFGS Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only)

  • 11 → Delocalized internal coordinates with BFGS using total energy Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only)

  • 12 → Isokinetic ensemble molecular dynamics. The equation of motion of the ions in contact with a thermostat are solved with the algorithm proposed in [Zhang1997], as worked out in [Minary2003]. The conservation of the kinetic energy is obtained within machine precision at each step. As in [Evans1983], when there is no fixing of atoms, the number of degrees of freedom in which the microscopic kinetic energy is hosted is 3*natom - 4. Indeed, the total kinetic energy is constrained, which accounts for minus one degree of freedom (also mentioned in [Minary2003]), but also there are three degrees of freedom related to the total momentum in each direction, that cannot be counted as microscopic degrees of freedom, since the total momentum is also preserved (but this is not mentioned in [Minary2003]). When some atom is fixed in one or more direction, e.g. using natfix, natfixx, natfixy, or natfixz, the number of degrees of freedom is decreased accordingly, albeit taking into account that the total momentum is not preserved anymore (e.g. fixing the position of one atom gives 3*natom-4, like in the non-fixed case). Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only) Related variables: time step (dtion) and the first temperature in mdtemp in case the velocities vel are not initialized, or all initialized to zero.

  • 13 → Isothermal/isenthalpic ensemble. The equation of motion of the ions in contact with a thermostat and a barostat are solved with the algorithm proposed in [Martyna1996]. If optcell=1 or 2, the mass of the barostat (bmass) must be given in addition. Purpose: Molecular dynamics Cell optimization: Yes (if optcell/=0) Related variables: The time step (dtion), the temperatures (mdtemp), the number of thermostats (nnos), and the masses of thermostats (qmass).

  • 14 → Simple molecular dynamics with a symplectic algorithm proposed in [Blanes2002] (called SRKNa14] of the kind first published in [Yoshida1990]. This algorithm requires at least 14 evaluation of the forces (actually 15 are done within Abinit) per time step. At this cost it usually gives much better energy conservation than the verlet algorithm (ionmov 6) for a 30 times bigger value of dtion. Notice that the potential energy of the initial atomic configuration is never evaluated using this algorithm. Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only)

  • 15 → Fast inertial relaxation engine (FIRE) algorithm proposed by Erik Bitzek, Pekka Koskinen, Franz Gähler, Michael Moseler, and Peter Gumbsch in [Bitzek2006]. According to the authors, the efficiency of this method is nearly the same as L-bfgs (ionmov = 22). It is based on conventional molecular dynamics with additional velocity modifications and adaptive time steps. The initial time step is set with dtion. Note that the physical meaning and unit of dtion are different from the default ones. The purpose of this algorithm is relaxation, not molecular dynamics. dtion governs the ion position changes, but the cell parameter changes as well. The positions are in reduced coordinates instead of in cartesian coordinates. The suggested first guess of dtion is 0.03. Purpose: Relaxation Cell optimization: Yes (if optcell/=0) Related variables: The initial time step dtion

  • 20 → Direct inversion of the iterative subspace. Given a starting point xred that is a vector of length 3*natom (reduced nuclei coordinates), and unit cell parameters (%rprimd) this routine uses the DIIS (direct inversion of the iterative subspace) to minimize the gradient (forces) on atoms. The preconditioning used to compute errors from gradients is using an inverse hessian matrix obtained by a BFGS algorithm. This method is known to converge to the nearest point where gradients vanish. This is efficient to refine positions around a saddle point for instance. Purpose: Structural optimization Cell optimization: No (Use optcell = 0 only) Related variables: DIIS memory diismemory

  • 22 → Conduct structural optimization using the Limited-memory Broyden-Fletcher-Goldfarb-Shanno minimization (L-BFGS) [Nocedal1980]. The routines are based on the original implementation by J. Nocedal available on netlib.org. This algorithm can be much better than the native implementation of BFGS in ABINIT (ionmov = 2) when one approaches convergence, perhaps because of better treatment of numerical details. Purpose: Structural optimization Cell optimization: Yes (if optcell/=0)

  • 23 → Use of Learn on The Fly method (LOTF) for Molecular Dynamics. In the framework of isokinetic MD, the atomic forces and positions are computed by using LOTF interpolation. A SCF computation is performed only any lotf_nitex steps. The results of the SCF are used to compute the parameters of a short range classical potential (for the moment only the glue potential for gold is implemented). Then these parameters are continuously tuned to compute atomic trajectories. LOTF has to be enabled at configure time. If LOTF is not enabled and ionmov = 23, abinit will set automatically ionmov = 12. The LOTF cycle is divided in the following steps: a) Initialization (SFC at t=0) and computation of potential parameters. b) Extrapolation of the atomic forces and positions for lotf_nitex time step. To perform this extrapolation, the potential computed in a) is used (Verlet algorithm). c) SFC at t=lotf_nitex. Computation of the potential parameters. d) LOTF interpolation, linear interpolation of the potential parameters and computation of the atomic forces and positions between t=0 and t=lotf_nitex. Purpose: Molecular Dynamics Cell optimization: No (Use optcell = 0 only) Related variables: dtion, lotf_classic, lotf_nitex, lotf_nneigx, lotf_version.

  • 24 → Simple constant energy molecular dynamics using the velocity Verlet symplectic algorithm (second order), see [Hairer2003]. The only related parameter is the time step (dtion). Purpose: Molecular dynamics Cell optimization: No (Use optcell = 0 only) Related variables: time step dtion

  • 25 → Hybrid Monte Carlo sampling of the ionic positions at fixed temperature and unit cell geometry (NVT ensemble). The underlying molecular dynamics corresponds to ionmov = 24. The related parameters are the time step (dtion) and thermostat temperature (mdtemp). Within the HMC algorithm [Duane1987], the trial states are generated via short NVE trajectories (ten ionmov = 24 steps in current implementation). The initial momenta for each trial are randomly sampled from Boltzmann distribution, and the final trajectory state is either accepted or rejected based on the Metropolis criterion. Such strategy allows to simultaneously update all reduced coordinates, achieve higher acceptance ratio than classical Metropolis Monte Carlo and better sampling efficiency for shallow energy landscapes [Prokhorenko2018]. Purpose: Monte Carlo sampling Cell optimization: No (Use optcell = 0 only) Related variables: time step dtion, thermostat temperature mdtemp,

  • 27 → TO BE DOCUMENTED

  • 28 → Update atomic positions and unit cell parameters using the i-pi client-server protocol as described in [Kapil2019]. In this case, ABINIT (the client) sends total energies, forces and stress tensor to the server that sends back the new configuration (atomic positions and cell parameters) to the ab-initio engine for further calculations. Note that the server is supposed to generate an input file that is consistent with the optimization/MD algorithm implemented by ABINIT. For instance, a server that wants to performs calculations with varying unit cells should set optcell > 0 in the initial input. Note that, at present, this feature is mainly used to interface ABINIT with the ASE optimization routines. Moreover the user is responsible for creating an input file with tuned tolerances to prevent Abinit from exiting when internal convergence is reached. See examples available in the ASE documentation Purpose: Structural optimization driver by the server (MD runs are not yet supported) Cell optimization: Yes (provide optcell > 0 in the initial input)

istatimg

Mnemonics: Integer governing the computation of STATic IMaGes
Mentioned in topic(s): topic_PIMD
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

This input variable is relevant when sets of images are activated (see imgmov). Not all images might be required to evolve from one time step to the other (seedynimage): these are static images. If istatimg = 0, the total energy of static images is not computed (but static images are used to make the dynamic images evolve). This can be useful to save CPU time. If istatimg = 1, the total energy of static images is computed.

mdtemp

Mnemonics: Molecular Dynamics TEMPeratures
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics
Variable type: real
Dimensions: (2)
Default value: [300, 300]
Added in version: before_v9

Test list (click to open). Moderately used, [13/1245] in all abinit tests, [1/159] in abinit tutorials

Give the initial and final temperature of the Nose-Hoover thermostat (ionmov = 8) and Langevin dynamics (ionmov = 9), in Kelvin. This temperature will change linearly from the initial temperature mdtemp(1) at itime=1 to the final temperature mdtemp(2) at the end of the ntime timesteps.

In the case of the isokinetic molecular dynamics (ionmov = 12), mdtemp(1) allows ABINIT to generate velocities (vel) to start the run if they are not provided by the user or if they all vanish. However mdtemp(2) is not used (even if it must be defined to please the parser). If some velocities are non-zero, mdtemp is not used, the kinetic energy computed from the velocities is kept constant during the run.

mdwall

Mnemonics: Molecular Dynamics WALL location
Mentioned in topic(s): topic_MolecularDynamics
Variable type: real
Dimensions: scalar
Default value: 10000.0
Comment: the walls are extremely far away
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the location (atomic units) of walls on which the atoms will bounce back when ionmov = 6, 7, 8 or 9. For each cartesian direction idir=1, 2 or 3, there is a pair of walls with coordinates xcart(idir)=-wall and xcart(idir)=rprimd(idir,idir)+wall. Supposing the particle will cross the wall, its velocity normal to the wall is reversed, so that it bounces back. By default, given in Bohr atomic units (1 Bohr=0.5291772108 Angstroms), although Angstrom can be specified, if preferred, since mdwall has the LENGTH characteristics.

mep_mxstep

Mnemonics: Minimal Energy Path search: MaXimum allowed STEP size
Characteristics: LENGTH
Mentioned in topic(s): topic_TransPath
Variable type: real
Dimensions: scalar
Default value: 0.4 if imgmov == 5, 100.0 otherwise.

Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Relevant only when imgmov = 1 (Steepest-Descent), 2 (String Method) or 5 (Nudged Elastic Band). The optimizer used to solve the Ordinary Differential Equation (ODE) can be constrained with a maximum allowed step size for each image. By default this feature is only activated for Nudged Elastic Band (NEB) and the value is inspired by [Sheppard2008]. Note that the step size is defined for each image as step = SQRT[SUM(R_i dot R_i)] where the R_i are the positions of the atoms in the cell.

mep_solver

Mnemonics: Minimal Energy Path ordinary differential equation SOLVER
Mentioned in topic(s): topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: None
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Relevant only when imgmov = 2 (String Method) or 5 (Nudged Elastic Band). Gives the algorithm used to solve the Ordinary Differential Equation (ODE) when searching for a Minimal Energy Path (MEP). Possible values can be:

  • 0 → Steepest-Descent algorithm following the (scaled) forces, the scaling factor being fxcartfactor (forward Euler method). Compatible with all MEP search methods.

  • 1 → Quick-min optimizer following the (scaled) forces, the scaling factor being fxcartfactor. The “quick minimizer” improves upon the steepest-descent method by accelerating the system in the direction of the forces. The velocity (of the image) is projected long the force and cancelled if antiparallel to it. Compatible only with Nudged Elastic Band (imgmov = 5). See [Sheppard2008].

  • 2 → Local Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm; each image along the band is minimized with a different instance of the BFGS optimizer. Compatible only with Nudged Elastic Band (imgmov = 5). See [Sheppard2008]. IN DEVELOPPMENT - NOT RELIABLE

  • 3 → Global Broyden-Fletcher-Goldfarb-Shanno (GL-BFGS) algorithm; all images along the band are minimized with a single instance of the BFGS optimizer. Compatible only with Nudged Elastic Band (imgmov = 5). See [Sheppard2008]. IN DEVELOPPMENT - NOT RELIABLE

  • 4 → Fourth-order Runge-Kutta method; the images along the band are moved every four steps (1 <=istep<=ntimimage) following the Runge-Kutta algorithm, the time step being fxcartfactor. Compatible only with Simplified String Method (imgmov = 2 and string_algo = 1 or 2). See [Weinan2007].

All of the optimizers can be constrained with a maximum allowed step size for each image; see mep_mxstep. This is by default the case of the Nudged Elastic Band (imgmov = 5).

mixesimgf

Mnemonics: MIXing Electronic Structure IMAGE Factors
Mentioned in topic(s): topic_CrossingBarriers
Variable type: real
Dimensions: (nimage)
Default value: None
Added in version: before_v9

Test list (click to open). Rarely used, [5/1245] in all abinit tests, [3/159] in abinit tutorials

Used in the algorithm Linear Combination of Constrained DFT Energies, that is, when imgmov==6.

This array gives, for each one of the nimage images, the factor by which the total energies for systems with same geometry but different electronic structures (occupation numbers) are linearly combined. The sum of these factors must equal 1.

natcon

Mnemonics: Number of AToms in CONstraint equations
Characteristics: NO_MULTI, INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: (nconeq)
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of atoms appearing in each of the nconeq independent equations constraining the motion of atoms during structural optimization or molecular dynamics (see nconeq, iatcon, and wtatcon).

natfix

Mnemonics: Number of Atoms that are FIXed
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: scalar
Default value: 0
Comment: (no atoms held fixed)
Added in version: before_v9

Test list (click to open). Moderately used, [33/1245] in all abinit tests, [6/159] in abinit tutorials

Gives the number of atoms (not to exceed natom) which are to be held fixed during a structural optimization or molecular dynamics. When natfix > 0, natfix entries should be provided in array iatfix .

natfixx

Mnemonics: Number of Atoms that are FIXed along the X direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of atoms (not to exceed natom) which are to be held fixed along the X direction during a structural optimization or molecular dynamics. When natfixx > 0, natfixx entries should be provided in array iatfixx.

natfixy

Mnemonics: Number of Atoms that are FIXed along the Y direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of atoms (not to exceed natom) which are to be held fixed along the Y direction during a structural optimization or molecular dynamics. When natfixy > 0, natfixy entries should be provided in array iatfixy

natfixz

Mnemonics: Number of Atoms that are FIXed along the Z direction
Characteristics: INPUT_ONLY
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of atoms (not to exceed natom) which are to be held fixed along the Z direction during a structural optimization or molecular dynamics. When natfixz > 0, natfixz entries should be provided in array iatfixz.

nconeq

Mnemonics: Number of CONstraint EQuations
Characteristics: NO_MULTI
Mentioned in topic(s): topic_GeoConstraints
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of independent equations constraining the motion of atoms during structural optimization or molecular dynamics (see natcon, iatcon, and wtatcon).

neb_algo

Mnemonics: Nudged Elastic Band ALGOrithm
Mentioned in topic(s): topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 1
Only relevant if: imgmov == 5
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the variant of the NEB method used. Possible values can be:

  • 0 → Original NEB method. See [Berne1998] pp. 385-404

  • 1 → NEB + improved tangent. The Improved Tangent Method builds on the NEB with an improved estimate of the tangent direction and a resulting change of the component of the spring force acting on the images. See [Henkelman2000]

  • 2 → Climbing-Image NEB (CI-NEB). The CI-NEB method constitutes a small modification to the NEB method. Information about the shape of the MEP is retained, but a rigorous convergence to a saddle point is also obtained. By default the spring constants are variable (see neb_spring). As the image with the highest energy has to be identified, the calculation begins with several iterations of the standard NEB algorithm. The effective CI-NEB begins at the cineb_start iteration. See [Henkelman2000a]

Note that, in all cases, it is possible to define the value of the spring constant connecting images with neb_spring, keeping it constant or allowing it to vary between 2 values (to have higher resolution close to the saddle point).

neb_spring

Mnemonics: Nudged Elastic Band: SPRING constant
Mentioned in topic(s): topic_TransPath
Variable type: real
Dimensions: (2)
Default value: [0.02, 0.15] Hartree/Bohr^2 if neb_algo == 2, [0.05, 0.05] Hartree/Bohr^2 otherwise.

Only relevant if: imgmov == 5
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the minimal and maximal values (in Hartree/Bohr^2) of the spring constant connecting images for the NEB method. In the standard “Nudged Elastic Band” method, the spring constant is constant along the path, but, in order to have higher resolution close to the saddle point, it can be better to have stronger springs close to it. See [Henkelman2000a]

nimage

Mnemonics: Number of IMAGEs
Mentioned in topic(s): topic_PIMD, topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Moderately used, [22/1245] in all abinit tests, [5/159] in abinit tutorials

Give the number of images (or replicas) of the system, for which the forces and stresses might be computed independently, in the context of string or NEB method, genetic algorithm, hyperdynamics, Path-Integral Molecular Dynamics, linear combination of images, pSIC, genetic algorithm, etc, depending on the value of imgmov). Related input variables: dynimage, npimage, ntimimage and prtvolimg. If nimage >1, the default choice for printing many files is set to zero, and the user might want to manually reestablish the printing, using, e.g. prtgsr, prtwf, prtebands, prteig, etc.

Images might differ by the position of atoms in the unit cell, their velocity, as well as by their cell geometry. The following input variables might be used to define the images:

These input variables, non-modified, will be used to define the image with index 1. For the image with the last index, the input file might specify the values of such input variables, appended with “_lastimg”, e.g.:

  • acell_lastimg
  • rprim_lastimg
  • xcart_lastimg
  • …

By default, these values will be interpolated linearly to define values for the other images, unless there exist specific values for some images, for which the string “last” has to be replaced by the index of the image, e.g. for the image number 4:

  • acell_4img
  • rprim_4img
  • xcart_4img
  • …

It is notably possible to specify the starting point and the end point of the path (of images), while specifying intermediate points.

It usually happen that the images do not have the same symmetries and space group. ABINIT has not been designed to use different set of symmetries for different images. ABINIT will use the symmetry and space group of the image number 2, that is expected to have a low number of symmetries. This might lead to erroneous calculations, in case some image has even less symmetry. By contrast, there is no problem if some other image has more symmetries than those of the second image.

nnos

Mnemonics: Number of NOSe masses
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the number of thermostats in the chain of oscillators thermostats as proposed in [Martyna1996]. The thermostat chains can be used either to perform Molecular Dynamics (MD) (ionmov = 13) or to perform Path Integral Molecular Dynamics (PIMD) (imgmov = 13). The mass of these thermostats is given by qmass.

noseinert

Mnemonics: NOSE thermostat INERTia factor
Mentioned in topic(s): topic_MolecularDynamics
Variable type: real
Dimensions: scalar
Default value: 100000
Only relevant if: ionmov == 8
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Give the inertia factor WT of the Nose-Hoover thermostat (when ionmov = 8), in atomic units of weight*length2, that is (electron mass)*(Bohr)2. The equations of motion are: MI d2RI/dt2= FI - dX/dt MI dRI/dt and WT d2X/dt2= Sum(I) MI (dRI/dt)2 - 3NkBT where I represent each nucleus, MI is the mass of each nucleus (see amu), RI is the coordinate of each nucleus (see xcart), dX/dt is a dynamical friction coefficient, and T is the temperature of the thermostat (see mdtemp).

ntime

Mnemonics: Number of TIME steps
Mentioned in topic(s): topic_MolecularDynamics, topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0 if ionmvov == 0, set to 1000 if ionvmov != 0 and imgmov != 0 and the variable is not specified.
Added in version: before_v9

Test list (click to open). Moderately used, [148/1245] in all abinit tests, [21/159] in abinit tutorials

Gives the maximum number of molecular dynamics time steps or structural optimization steps to be done if ionmov is non-zero. Starting with Abinit9, ntime is automatically set to 1000 if ionmov is non-zero, ntimimage is zero and ntime is not specified in the input file. Users are encouraged to pass a timelimit to Abinit using the command line and the syntax:

    abinit --timelimit hours:minutes:seconds

so that the code will try to stop smoothly before the timelimit and produce the DEN and the WFK files that may be used to restart the calculation.

Note that at the present the option ionmov = 1 is initialized with four Runge-Kutta steps which costs some overhead in the startup. By contrast, the initialisation of other ionmov values is only one SCF call. Note that ntime is ignored if ionmov = 0.

ntimimage

Mnemonics: Number of TIME steps for IMAGE propagation
Mentioned in topic(s): topic_PIMD, topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Moderately used, [21/1245] in all abinit tests, [5/159] in abinit tutorials

Gives the maximal number of molecular dynamics time steps or structural optimization steps to be done for the set of images, referred to as ‘image- timesteps’. At each image-timestep, all the images are propagated simultaneously, each according to the algorithm determined by imgmov and the usual accompanying input variables, and then the next positions and velocities for each image are determined from the set of results obtained for all images.

optcell

Mnemonics: OPTimize the CELL shape and dimensions
Mentioned in topic(s): topic_PIMD, topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Moderately used, [68/1245] in all abinit tests, [10/159] in abinit tutorials

Allows one to optimize the unit cell shape and dimensions, when ionmov >= 2 or 3. The configuration for which the stress almost vanishes is iteratively determined, by using the same algorithms as for the nuclei positions. May modify acell and/or rprim. The ionic positions are ALWAYS updated, according to the forces. A target stress tensor might be defined, see strtarget. In most cell relaxations ( optcell /=0), the user should define ecutsm, dilatmx, and tolrff, because the default values are NOT aimed at cell relaxation runs. The user should also be aware that convergence studies might need to modify the value of tolmxf, the default being decent, but possibly not sufficiently stringent.

  • optcell = 0: modify nuclear positions, since ionmov = 2 or 3, but no cell shape and dimension optimisation.
  • optcell = 1: optimisation of volume only (do not modify rprim, and allow an homogeneous dilatation of the three components of acell)
  • optcell = 2: full optimization of cell geometry (modify acell and rprim - normalize the vectors of rprim to generate the acell). This is the usual mode for cell shape and volume optimization. It takes into account the symmetry of the system, so that only the effectively relevant degrees of freedom are optimized.
  • optcell = 3: constant-volume optimization of cell geometry (modify acell and rprim under constraint - normalize the vectors of rprim to generate the acell)
  • optcell = 4, 5 or 6: optimize (both length and angle) the first, second or third vector, respectively (only works if, in the case optcell=6, the two other vectors lie in the xy cartesian plane; similarly for optcell=4 and optcell=5, the two clamped vectors must lie in the yz and xz plane respectively).
  • optcell = 7, 8 or 9: optimize the cell geometry while keeping the first, second or third vector unchanged (only works if the two other vectors are orthogonal to the one left unchanged, the latter being along its cartesian axis).

A few details require attention when performing unit cell optimisation:

  • one has to get rid of the discontinuities due to discrete changes of plane wave number with cell size, by using a suitable value of ecutsm;
  • one has to allow for the possibility of a larger sphere of plane waves, by using dilatmx;
  • one might have to adjust the scale of stresses to the scale of forces, by using strfact.
  • if all the reduced coordinates of atoms are fixed by symmetry, one cannot use tolrff (or toldff) to stop the SCF cycle. (Suggestion: use tolvrs with a small value, like 1.0d-14)

It is STRONGLY suggested first to optimize the ionic positions without cell shape and size optimization (optcell = 0), then start the cell shape and size optimization from the cell with relaxed ionic positions, possibly from the previous dataset, using getxred or getxcart. Presently (v3.1), one cannot restart (restartxf) a calculation with a non-zero optcell value from the (x,f) history of another run with a different non- zero optcell value. There are still a few problems at that level.

pimass

Mnemonics: Path Integral fictitious MASSes
Mentioned in topic(s): topic_PIMD
Variable type: real
Dimensions: (ntypat)
Default value: ntypat
Only relevant if: imgmov = 9 or 13
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Only relevant if imgmov = 9 or 13 (Path-Integral Molecular Dynamics). Gives the fictitious masses ( [Marx1996]) in atomic mass units for each kind of atom in cell. These masses are the inertial masses used in performing Path Integral Molecular Dynamics (PIMD), they are different from the true masses (amu) used to define the quantum spring that relates the different beads in PIMD. They can be chosen arbitrarily, but an appropriate choice will lead the different variables to move on the same time scale in order to optimize the sampling efficiency of the PIMD trajectory. If pitransform = 1 (normal mode transformation), or pitransform = 2 (staging transformation), pimass is automatically set to its optimal value.

pimd_constraint

Mnemonics: Path-Integral Molecular Dynamics: CONSTRAINT to be applied on a reaction coordinate
Mentioned in topic(s): topic_PIMD
Variable type: integer
Dimensions: scalar
Default value: 0
Only relevant if: imgmov = 9 or 13
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Only relevant for Path-Integral Molecular Dynamics. Selects a constraint to be applied during the PIMD trajectory. The constraint is holonomic (it is a relation between the position variables). In practice, the total forces applied to the atomic positions are modified so as to respect the constraint.

To date, the available constraints are:

  • 0: no constraint
  • 1: “Blue Moon Ensemble” method. The constraint is a linear combination of the positions of atomic centroids (this linear combination is kept constant during the simulation). Sum[W_i * X_i] = constant The X_i are the coordinates of the atomic centroids. The weights W_i have to be specified with the wtatcon(3,natcon,nconeq), iatcon(natcon) and natcon input parameters (where nconeq is fixed to 1). More details on the implementation in [Komeiji2007].

pitransform

Mnemonics: Path Integral coordinate TRANSFORMation
Mentioned in topic(s): topic_PIMD
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

Only relevant if imgmov = 9 or 13 (Path-Integral Molecular Dynamics). Coordinate transformation used in the integration of the Path Integral Molecular Dynamics equations of motion. The transformation, with an appropriate choice of fictitious masses (pimass), is used to force the different modes to move on the same time scale, and thus optimize the efficiency of the statistical sampling in the corresponding statistical ensemble. Available with a Langevin thermostat (imgmov = 9) or with Nose- Hoover chains (imgmov = 13). See [Tuckerman1996].

If equal to 0, no transformation is applied (primitive coordinates). If equal to 1, normal mode transformation (in that case, nimage must be absolutely EVEN). If equal to 2, staging transformation.

prtatlist

Mnemonics: PRinT by ATom LIST of ATom
Characteristics: NO_MULTI
Mentioned in topic(s): topic_printing, topic_Output
Variable type: integer
Dimensions: (natom)
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

This is an array of the numbers associated to the index atoms that the user want to print in the output or log files, this is useful when you have a large number of atoms and you are only interested to follow specific atoms, the numbers associated should be consistent with the list in xcart or xred. This input variable does not affect the contents of the “OUT.nc” or “HIST.nc”, those are NetCDF files containing the information about all the atoms.

qmass

Mnemonics: Q thermostat MASS
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics
Variable type: real
Dimensions: (nnos)
Default value: 10.0
*Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

This are the masses of the chains of nnos thermostats to be used when ionmov = 13 (Molecular Dynamics) or imgmov = 13 (Path Integral Molecular Dynamics).

If ionmov = 13 (Molecular Dynamics), this temperature control can be used with optcell =0, 1 (homogeneous cell deformation) or 2 (full cell deformation). If imgmov = 13 (Path Integral Molecular Dynamics), this temperature control can be used with optcell =0 (NVT ensemble) or 2 (fully flexible NPT ensemble). In that case, optcell = 2 is NOT USABLE yet.

random_atpos

Mnemonics: RANDOM ATomic POSitions
Mentioned in topic(s): topic_crystal, topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Control the inner coordinates, which can be generated randomly by using 4 different methods depending ont its value (0) if zero, no random generation and xred are taken as they have been introduced by the user (1) if one, particles are generated completely random within the unit cell. (2) if two, particles are generated randomly but the inner particle distance is always larger than a factor of the sum of the covalent bonds between the atoms (note: this is incompatible with the definition of alchemical mixing, in which ntypat differs from npsp)

restartxf

Mnemonics: RESTART from (X,F) history
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics, topic_GeoOpt
Variable type: integer
Dimensions: scalar
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [5/1245] in all abinit tests, [0/159] in abinit tutorials

Control the restart of a molecular dynamics or structural optimization job.

restartxf = 0 (Default): No restart procedure is enabled and the code will start a Molecular dynamics or structural optimization from scratch.

restartxf = -1 (New): Use the HIST.nc file to reconstruct a partial calculation. It will reconstruct the different configurations using the forces and the stresses stored in the HIST.nc file, instead of calling the SCF procedure. Using restartxf = -1 from the beginning is harmless. The only condition is to keep the input file the same in such a way that the same predictor is used and it will predict the same structure recorded in the HIST.nc file. This option will always compute extra ntime iterations independent of the number of iterations recovered previously.

restartxf = -2 (New): Read the HIST.nc file and select the atomic positions and cell parameters with the lowest energy. Forget all the history and start the calculation using those values. The original atomic coordinates and cell parameters are irrelevant in this case.

restartxf = -3 (New): Read ONLY the last required atomic positions and cell parameters in the HIST.nc file to restart the Molecular dynamics or structural optimization.

NOTES:

  • You can use restartxf=-1, -2 or -3 for all predictors that make no use of random numbers.
  • You can use restartxf=-1, -2 or -3 to restart a calculation that was not completed. The HIST.nc file is written on each iteration. So you always have something to recover from.
  • You can take advantage of the appropriate values of irdwfk or getwfk to get a good wave function to continue your job.

Previously, positive values of restartxf were allowed, but these are deprecated since ABINITv8.

signperm

Mnemonics: SIGN of PERMutation potential
Mentioned in topic(s): topic_MolecularDynamics
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [2/1245] in all abinit tests, [0/159] in abinit tutorials

+1 favors alternation of species -1 favors segregation

strfact

Mnemonics: STRess FACTor
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 100
Added in version: before_v9

Test list (click to open). Rarely used, [9/1245] in all abinit tests, [2/159] in abinit tutorials

The stresses (in atomic units) multiplied by strfact will be treated like forces (in atomic units) in the algorithms for optimization (ionmov = 2 or 22, non-zero optcell). For example, the stopping criterion defined by tolmxf relates to these scaled stresses.

string_algo

Mnemonics: STRING method ALGOrithm
Mentioned in topic(s): topic_TransPath
Variable type: integer
Dimensions: scalar
Default value: 1
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Relevant only when imgmov = 2 (String Method). Gives the variant of the String Method method used. Possible values can be:

  • 0 → Original String Method. NOT YET IMPLEMENTED. See [Weinan2002]

  • 1 → Simplified String Method with parametrization by equal arc length. Instead of using the normal force (wrt the band), the full force is used; the reparametrization is enforced by keeping the points of the string equally spaced. See [Weinan2007]

  • 2 → Simplified String Method with parametrization by energy-weighted arc length. A variant of the Simplified String Method (like 1-); the reparametrization is done by using energy-weight arc-lengths, giving a finer distribution near the saddle point. See [Weinan2007] and [Goodrow2009]

strprecon

Mnemonics: STRess PRECONditioner
Mentioned in topic(s): topic_ForcesStresses, topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 1.0
Added in version: before_v9

Test list (click to open). Rarely used, [4/1245] in all abinit tests, [0/159] in abinit tutorials

This is a scaling factor to initialize the part of the Hessian related to the treatment of the stresses (optimisation of the unit cell). In case there is an instability, decrease the default value, e.g. set it to 0.1.

strtarget

Mnemonics: STRess TARGET
Mentioned in topic(s): topic_ForcesStresses, topic_GeoOpt
Variable type: real
Dimensions: (6)
Default value: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Define the target stress to be obtained by optimization of the cell. The components of the stress tensor must be stored according to: (1,1) →1; (2,2) → 2; (3,3) → 3; (2,3) → 4; (3,1) → 5; (1,2) →6. The conversion factor between Ha/Bohr**3 (the atomic unit for stress) and GPa is: 1 Ha/Bohr**3 = 29421.033 GPa. If a hydrostatic stress (so, a pressure P) is wanted, define strtarget as -P -P -P 0 0 0 . Not used if optcell == 0.

tolimg

Mnemonics: TOLerance on the mean total energy for IMaGes
Characteristics: ENERGY
Mentioned in topic(s): topic_TransPath
Variable type: real
Dimensions: scalar
Default value: 5e-05
Added in version: before_v9

Test list (click to open). Moderately used, [16/1245] in all abinit tests, [5/159] in abinit tutorials

Sets a maximal absolute energy tolerance (in hartree, averaged over dynamic images) below which iterations on images (the one governed by the ntimimage input variable) will stop. This is to be used when trying to optimize a population of structures to their lowest energy configuration, taking into account the particular algorithm defined by imgmov A value of about 5.0d-5 hartree or smaller is suggested (this corresponds to about 3.7d-7 eV). No meaning for RF calculations.

tolmxde

Mnemonics: TOLerance on the MaXimal Difference in Energy
Characteristics: ENERGY
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 0.0
Added in version: before_v9

Test list (click to open). Rarely used, [1/1245] in all abinit tests, [0/159] in abinit tutorials

Sets a maximal difference in energy with respect to the two previous steps below which BFGS structural relaxation iterations will stop. A value of about 0.0005 eV/atom or smaller is suggested. In order to use tolmxde, you should explicitly set tolmxf to 0.0. No meaning for RF calculations.

tolmxf

Mnemonics: TOLerance on the MaXimal Force
Mentioned in topic(s): topic_GeoOpt
Variable type: real
Dimensions: scalar
Default value: 5e-05
Added in version: before_v9

Test list (click to open). Moderately used, [99/1245] in all abinit tests, [18/159] in abinit tutorials

Sets a maximal absolute force tolerance (in hartree/Bohr) below which BFGS structural relaxation iterations will stop. Can also control tolerance on stresses, when optcell /=0, using the conversion factor strfact. This tolerance applies to any particular cartesian component of any atom, excluding fixed ones. See the parameter ionmov. This is to be used when trying to equilibrate a structure to its lowest energy configuration. A value of about 5.0d-5 hartree/Bohr or smaller is suggested (this corresponds to about 2.5d-3 eV/Angstrom). No meaning for RF calculations.

vel

Mnemonics: VELocity
Characteristics: EVOLVING
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics
Variable type: real
Dimensions: (3,natom)
Commentdims: It is represented internally as vel(3,natom,nimage)
Default value: 0
*Only relevant if: ionmov > 0
Added in version: before_v9

Test list (click to open). Moderately used, [107/1245] in all abinit tests, [16/159] in abinit tutorials

Gives the starting velocities of atoms, in cartesian coordinates, in Bohr/atomic time units (atomic time units given where dtion is described). For ionmov = 8 (Nose thermostat), if vel is not initialized, a random initial velocity giving the right kinetic energy will be generated. If the atom manipulator is used, vel will be related to the preprocessed set of atoms, generated by the atom manipulator. The user must thus foresee the effect of this atom manipulator (see objarf). Velocities evolve is ionmov == 1.

vel_cell

Mnemonics: VELocity of the CELL parameters
Characteristics: EVOLVING
Mentioned in topic(s): topic_PIMD
Variable type: real
Dimensions: (3,3)
Commentdims: It is represented internally as vel_cell(3,3,nimage)
Default value: 3
*Only relevant if: imgmov in [9,13] and optcell > 0 (Path-Integral Molecular Dynamics with NPT algorithm)
Added in version: before_v9

Test list (click to open). Rarely used, [0/1245] in all abinit tests, [0/159] in abinit tutorials

Irrelevant unless imgmov = 9 or 13 and optcell>0 (Path-Integral Molecular Dynamics with NPT algorithm). Gives the starting velocities of the dimensional cell parameters in Bohr/atomic time units (atomic time units given where dtion is described).

vis

Mnemonics: VIScosity
Mentioned in topic(s): topic_PIMD, topic_MolecularDynamics
Variable type: real
Dimensions: scalar
Default value: 100
Added in version: before_v9

Test list (click to open). Moderately used, [24/1245] in all abinit tests, [3/159] in abinit tutorials

The equation of motion is:

M I d 2 R I /dt 2 = F I - vis dR I /dt

The atomic unit of viscosity is hartree * (atomic time units)/Bohr 2. Units are not critical as this is a fictitious damping used to relax structures. A typical value for silicon is 400 with dtion of 350 and atomic mass 28 amu. Critical damping is most desirable and is found only by optimizing vis for a given situation.

In the case of Path-Integral Molecular Dynamics using the Langevin Thermostat (imgmov = 9), vis defines the friction coefficient, in atomic units. Typical value range is 0.00001-0.001.

wtatcon

Mnemonics: WeighTs for AToms in CONstraint equations
Characteristics: NO_MULTI
Mentioned in topic(s): topic_GeoConstraints
Variable type: real
Dimensions: (3,natcon,nconeq)
Default value: 0
Added in version: before_v9

Test list (click to open). Rarely used, [3/1245] in all abinit tests, [0/159] in abinit tutorials

Gives the weights determining how the motion of atoms is constrained during structural optimization or molecular dynamics (see nconeq, natcon, and iatcon). For each of the nconeq independent constraint equations, wtatcon is a 3*natcon array giving weights, W I, for the x, y, and z components of each of the atoms (labeled by I) in the list of indices iatcon. Prior to taking an atomic step, the calculated forces, F I, are replaced by projected forces, F’ I, which satisfy the set of constraint equations

Sum mu=x,y,z; I=1,natcon: W mu,I * F’ mu,I = 0 for each of the nconeq arrays W I.

Different types of motion constraints can be implemented this way. For example,

nconeq 1 natcon 2 iatcon 1 2 wtatcon 0 0 +1 0 0 -1

could be used to constrain the relative height difference of two adsorbate atoms on a surface (assuming their masses are equal), since F’ z,1 - F’ z,2 = 0 implies z 1 - z 2 = constant.