Variables: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
|||
| Line 131: | Line 131: | ||
{{var-basic|CUTGeo|None|String}} | {{var-basic|CUTGeo|None|String}} | ||
{{var-short|Cutoff geometry}} | {{var-short|Cutoff geometry}} | ||
{{var-usage|Allowed values are: "box/cylinder/sphere/ws X/Y/Z/XY...", e.g. "box xy" or "cylinder y". WS is the suggested option for orthorhombic cells. Use sphere (0D) for molecules, cylinder (1D) for polymers and nanotubes, box (0D, 1D, 2D) for all geometries. XYZ: cut in all directions. Box: XY: cut in XY only. Cylinder X/Y/Z indicates cylinder axis. When using Box shapes, the RIM is also needed to calculate the potential. In Box for large enough boxes assigns Box side slighlty smaller than the cell box}} ''See also:'' <code>[[Variables#CUTBox/CUTRadius/CUTCylLen/CUTwsGvec| | {{var-usage|Allowed values are: "box/cylinder/sphere/ws X/Y/Z/XY...", e.g. "box xy" or "cylinder y". WS is the suggested option for orthorhombic cells. Use sphere (0D) for molecules, cylinder (1D) for polymers and nanotubes, box (0D, 1D, 2D) for all geometries. XYZ: cut in all directions. Box: XY: cut in XY only. Cylinder X/Y/Z indicates cylinder axis. When using Box shapes, the RIM is also needed to calculate the potential. In Box for large enough boxes assigns Box side slighlty smaller than the cell box}} ''See also:'' <code>[[Variables#CUTBox|CUTBox]]</code><code>[[Variables#CUTRadius|CUTRadius]]</code><code>[[Variables#CUTCylLen|CUTCylLen]]</code><code>[[Variables#CUTwsGvec|CUTwsGvec]]</code> | ||
<span id=CUTBox></span> | <span id=CUTBox></span> | ||
{{var-basic|CUTBox|Length|Vector}} | {{var-basic|CUTBox|Length|Vector}} | ||
| Line 168: | Line 168: | ||
{{var-short|Provides in oututput the truncated Coulomb potential in real space}} | {{var-short|Provides in oututput the truncated Coulomb potential in real space}} | ||
{{var-usage|Useful option for debugging purpose}} | {{var-usage|Useful option for debugging purpose}} | ||
===Hartree-Fock Self-energy and Vxc=== | ===Hartree-Fock Self-energy and Vxc=== | ||
<span id=HF_and_locXC></span> | <span id=HF_and_locXC></span> | ||
Revision as of 12:09, 6 November 2019
This page gives a brief explanation of the many input variables used by Yambo.
Variables apply to the standard yambo executable unless indicated otherwise.
Default values are read from the indicated database file unless stated otherwise.
Required verbosity flags are indicated where appropriate.
Units can have the following values:
- None = no unit
- RL = number of G-vectors
- Energy = Ha, mHa (Hartree atomic units), Ry, mRy (Rydberg atomic units), eV, meV (electron volts), K, THz, GHz
- Length = Bohr atomic units a.u.
Formats are described on the Format page.
DO NOT EDIT THIS PAGE DIRECTLY!!! CHANGES WILL BE LOST!!!
All Variables (alphabetical order)
BDmRange BEnRange BEnSteps BLongDir BndsRnXd BndsRnXp BoseCut BoseTemp
BS_CPU BS_nCPU_LinAlg_DIAGO BS_nCPU_LinAlg_INV BS_ROLEs bse BSEBands BSEmod BSENGBlk
BSENGexx BSEPSInvTrs BSHayTrs bsk BSKmod BSSInvMode BSSInvPFratio BSSmod
CGrdSpXd chi ChiLinAlgMod Chimod cohsex CUTBox CUTCol_test CUTCylLen
CUTGeo CUTRadius CUTwsGvec DbGdQsize DbGdQsize DBsFRAGpm DBsIOoff DIP_Threads
DipApproach DipApproach DipPDirect DipPDirect DmERefXd DmRngeXd DmRngeXp DrClassic
DrudeWBS DrudeWXd dScStep DysSolver EhEngyXd ElecTemp Em1Anys em1d
em1s EMStpsXd EnRngeXd EnRngeXp ETStpsXd ETStpsXp EvalCurrent ExtendOut
ExtF_Dir ExtF_kind EXXRLvcs EXXRLvcs FFTGvecs FxcGRLc Gauge GbndRnge
GDamping GDmRnge GEnRnge GEnSteps GfnQP_E GfnQP_N GfnQP_Wc GfnQP_Wc_dos
GfnQP_Wc_E GfnQP_Wv GfnQP_Wv_dos GfnQP_Wv_E GfnQP_Z GfnQPdb GrFnTpXd GTermEn
GTermKind gw0 HARRLvcs HF_and_locXC IDEm1Ref IkSigLim IkXLim K_Threads
KfnQP_E KfnQP_N KfnQP_Wc KfnQP_Wc_dos KfnQP_Wc_E KfnQP_Wv KfnQP_Wv_dos KfnQP_Wv_E
KfnQP_Z KfnQPdb LifeTrCG LongDrXd LongDrXp LRC_alpha MaxGvecs MEM_tresh
MetDamp Nelectro NewtDchk NGsBlkXd NLBands NLCorrelation NLDamping NLintegrator
NLogCPUs nloptics NLstep NLtime NLverbosity NoCondSumRule NonPDirs OccTresh
OnMassShell optics PAR_def_mode PAR_def_mode PPAPntXp Qdirection QPerange QPerange
QpgFull QPkrange QPkrange QpntsRXd QptCoord QShiftOrder RandGvec RandQpts
Reflectivity rim_cut SE_CPU SE_ROLEs SE_Threads setup ShiftedPaths ShiftedPaths
StdoHash UseNLCC VXCRLvcs WehCpl WFbuffIO X_all_q_CPU X_all_q_nCPU_LinAlg_INV X_all_q_ROLEs
X_finite_q_CPU X_finite_q_nCPU_LinAlg_INV X_finite_q_ROLEs X_q_0_CPU X_q_0_nCPU_LinAlg_INV X_q_0_ROLEs X_Threads XfnQP_E
XfnQP_N XfnQP_Wc XfnQP_Wc_dos XfnQP_Wc_E XfnQP_Wv XfnQP_Wv_dos XfnQP_Wv_E XfnQP_Z
XfnQPdb
Global options
Nelectro(Units: None, Format: )Default: From ns.db1 Verbosity:-V gen- Meaning: Electrons number
- Usage:
StdoHash(Units: None, Format: Integer)Default: 40 Verbosity:-V io- Meaning: Number of hashes in live-timing output.
- Usage: Might be useful to increase for very long jobs, or if the code is hanging.
DBsIOoff(Units: None, Format: String)Default: None Verbosity:-V io- Meaning: List of databases not written to disk
- Usage: Space-separated list of DB with NO I/O. DB is (DIP,X,HF,COLLs,J,GF,CARRIERs,OBS,W,SC,BS,ALL). No ndb.* file is written.
- Example: DBsIOoff= "DIP" means ndb.dip_iR_and_P_fragment_* is not written, but stored in memory if Yambo needs it.
DBsFRAGpm(Units: None, Format: String)Default: None Verbosity:-V io- Meaning: List of databases to be fragmented
- Usage: Space-separated list of +DB to FRAG and -DB to NOT FRAG, where DB is (DIP,X,W,HF,COLLS,K,BS,QINDX,RT,ELP. Fragments the database. Smaller files (e.g. ndb.em1s_fragment_*) are created instead of a large one (e.g. ndb.em1s). Faster read/write operations in parallel runs
WFbuffIO(Units: None, Format: Flag)Default: Off Verbosity:-V io- Meaning: Wave-functions buffered I/O
- Usage: Parts of the WFs are stored by the node. Nodes communicate when these elements are needed. Memory heavy.
MEM_tresh(Units: None, Format: ) Verbosity:-V gen- Meaning: Threshold on traced memory allocations/deallocations
- Usage:
FFTGvecs(Units: RL/Energy, Format: Integer/Real) Verbosity:-V RL- Meaning: Number of G-vectors or energy cut off for expanding the wavefunctions/FFT transforms
- Usage: Determines size (memory) of calculation. Corresponds to cutoff in DFT calculation; can be much less than geometry cutoff. It needs careful convergence.
Initialization
MaxGvecs(Units: RL/Energy, Format: Integer/Real) Verbosity:-V RL- Meaning: Maximum number of G-vectors that can be used by code
- Usage:
IkSigLim(Units: Range, Format: Integer ) Verbosity:-V kpt- Meaning: QP K-points indices range
- Usage:
Random integration method and cutoff Coulomb potentials
RandQpts(Units: RL, Format: Integer)- Meaning: Number of random q-points in the BZ to perform Monte Carlo Integration,
- Usage: It needs convergence: values like 10^6 can be used to ensure convergence. Needed for non 3D system to avoid divergences for small q, and needed to build cutoff potential with box shape See also:
RandGvec
CUTGeo(Units: None, Format: String)- Meaning: Cutoff geometry
- Usage: Allowed values are: "box/cylinder/sphere/ws X/Y/Z/XY...", e.g. "box xy" or "cylinder y". WS is the suggested option for orthorhombic cells. Use sphere (0D) for molecules, cylinder (1D) for polymers and nanotubes, box (0D, 1D, 2D) for all geometries. XYZ: cut in all directions. Box: XY: cut in XY only. Cylinder X/Y/Z indicates cylinder axis. When using Box shapes, the RIM is also needed to calculate the potential. In Box for large enough boxes assigns Box side slighlty smaller than the cell box See also:
CUTBoxCUTRadiusCUTCylLenCUTwsGvec
CUTwsGvec(Units: Energy, Format: Real)- Meaning: Energy cut off on modified component
- Usage: {{{1}}} See also:
CUTGeo
RandGvec(Units: Energy, Format: ) Verbosity:-V RL- Meaning: Number of G vectors the RIM is calculated at
- Usage: {{{1}}}
CUTCol_test(Units: None, Format: Flag)- Meaning: Provides in oututput the truncated Coulomb potential in real space
- Usage: Useful option for debugging purpose
Hartree-Fock Self-energy and Vxc
EXXRLvcs(Units: RL/Energy, Format: Integer/Real)- Meaning: Number of G-vectors used in the sum of the exchange self-energy Sx.
- Usage: It needs careful convergence. As it is not particularly time consuming, large values can be used to ensure convergence. Generally a large number is needed as the QP energies show a slow convergence. The calcualtion of the exchange part is rather fast.
QPkrange(Units: None, Format: Range)- Meaning: Range of states (nk) where GW/Sx elements are calculated
- Usage: Careful use of fewer k-points and bands reduces the calculation time; yambo will interpolate the rest. Format is: first k-point See also:
QPerange
QPerange(Units: None/Energy, Format: Range) Verbosity:-V qp- Meaning:
- Usage: Careful use of fewer k-points and bands reduces the calculation time; yambo will interpolate the rest. Format is: first k-point See also:
QPkrange
SE_ROLEs(Units: None, Format: String) Verbosity:-V par- Meaning: CPUs roles (q,qp,b)
- Usage: {{{1}}}
SE_Threads(Units: None, Format: Integer) Verbosity:-V par- Meaning: Number of threads for self-energy
- Usage: Very efficient.
GW
QPkrange(Units: None, Format: Range)- Meaning: K-points and band range where you want to calculate the GW correction.
- Usage: If interested in non consecutive kpoints or bands multiple rows can be also considered. The syntax is first kpoint
GbndRnge(Units: None, Format: Range)- Meaning: Specifies the range of bands entering in the sum over states in the correlation part of the self energy
- Usage: It needs several empty states. Single quasiparticle states converges very slowly with respect GbndRnge, energy differences (e.g. gaps) behave better. See also GTermKind variable in order to speed up the convergences.This number is usually larger than the number of bands used to calculated the dielectricconstant. Single quasiparticle energies converge slowly with respect GbndRnge, energy difference behave better. You can use terminator technique to mitigate the slow dependence.
GDamping(Units: , Format: )- Meaning: Small damping in the Green's function definition, the delta parameter.
- Usage: The final result shouuld not depend on that, usually set at 0.1 eV
DysSolver(Units: , Format: )- Meaning: Indicates method used to solve the Dyson equation
- Usage: "n" Newton linearization (First order expansion around KS eigenvalue), 's' non linear iterative secant method
GfnQPdbSeeKfnQPdb
GfnQP_NSeeKfnQP_N
GfnQP_ESeeKfnQP_E
GfnQP_ZSeeKfnQP_Z
GfnQP_Wv_ESeeKfnQP_Wv_E
GfnQP_WvSeeKfnQP_Wv
GfnQP_Wv_dosSeeKfnQP_Wv_dos
GfnQP_Wc_ESeeKfnQP_Wc_E
GfnQP_WcSeeKfnQP_Wc
GfnQP_Wc_dosSeeKfnQP_Wc_dos
GTermKind(Units: None, Format: String) Verbosity:-V qp- Meaning: Type of terminator to accelarate onvergence with respect empty states
- Usage: Default is "none", possible options are "BG" for the Bruneval-Gonze terminator. See BG[1]. It speeds up the convergence with respect to number of empty bands.
ExtendOut(Units: None, Format: Flag) Verbosity:-V qp- Meaning: Extended output: Print more quantities in qp output files
- Usage:
Screening
NGsBlkXd(Units: RL/Energy, Format: Integer/Real)- Meaning: Number of G-vectors or energy cut off in the screening (response block size)
- Usage: Determines the size of the dielectric matrix in G-space. Needed for inclusion of local field effects, it should be much smaller than the number of G-vectors needed to expand the wavefunctions (FFTGvecs). It needs careful convergence, and should be converged along with BndsRnXp. Depends strongly on the system inhomogeneity. See also:
BndsRnXp FFTGvecs
PPAPntXp(Units: Energy, Format: Real)Default: 1 Ha (27.2114eV)- Meaning: Plasmon pole Imaginary Energy
- Usage: The self energy in the imaginary axis should be a smooth function so it should not have a strong dependence on this pole energy. Set it at an higher value of the plasmon energy (see EELS spectrum). this is the second frequency used to fit the Godby-Needs plasmon-pole model (PPM). If results depend consistently by changing this frequency, the PPM is not adequate for your calculation and it is need to gp beyond that, e.g. Real-axis.
BndsRnXpSeeBndsRnXd
EnRngeXpSeeEnRngeXd
DmRngeXpSeeDmRngeXd
ETStpsXpSeeETStpsXd
LongDrXpSeeLongDrXd
Optics/chi in G-space
Chimod(Units: None, Format: String)- Meaning: Type of kernel in TDDFT Dyson equation
- Usage: Do not set this manually - set using the appropriate -k command line option. Allowed values: IP/Hartree/ALDA/LRC/BSfxc. IP indicates no kernel (independent particle level/RPA without local fields)
X_Threads(Units: None, Format: Integer)- Meaning: Number of threads for response functions
- Usage: efficient, need extra mem
QpntsRXd(Units: None, Format: Range)- Meaning: Range of indexes of q-points/transferred momenta to be computed.
- Usage: Set to 1 1 to select just the long wavelength term. Range from 1:nqpt.
BndsRnXd(Units: None, Format: Range)- Meaning: Bands range: Specifies the number of bands entering in the sum over states in the RPA response function
- Usage: It needs several empty states. See also GTermKind variable in order to speed up the convergences. Reduce range in order to lower memory. In metals it includes partially filled bands. See also EhEngyXd. See also:
GTermKind
EnRngeXd(Units: Energy, Format: Range)- Meaning: Energy range the spectrum is calculated across.
- Usage: Extremae of the energy range across which optical spectra will be computed. See also:
EhEngyXd
DmRngeXd(Units: Energy, Format: Range)- Meaning: Determines the damping used across the requested spectral range.
- Usage: Typically this is kept constant. If different values are used, the damping at each energy will be interpolated linearly. This can be useful when poor k-point sampling leads to large oscillations at higher energy.
ETStpsXd(Units: None, Format: Integer)- Meaning: Number of energy steps in computing X
- Usage: Determines the number of steps in energy the response function/spectrum is computed for in the desired range defined by EnRngeXd. In the case of a full frequency GW the range is fixed by the occupied/empty states included in the calculation, the number of frequency requires a careful check
LongDrXd(Units: Length, Format: Vector)- Meaning: Electric field direction
- Usage: Pay attention how the system is oriented when treating non 3D systems and choose a direction in the plane/axis where your system lies
FxcGRLc(Units: , Format: )- Meaning: XC-kernel size
- Usage: Needs convergence study. Much less than FFTGvecs
LRC_alpha(Units: , Format: )- Meaning: LRC fitting parameter
- Usage: Long-range tail of the fxc kernel. Depends on the system: the larger the screening the smaller this parameter.
XfnQPdbSeeKfnQPdb
XfnQP_NSeeKfnQP_N
XfnQP_ESeeKfnQP_E
XfnQP_ZSeeKfnQP_Z
XfnQP_Wv_ESeeKfnQP_Wv_E
XfnQP_WvSeeKfnQP_Wv
XfnQP_Wv_dosSeeKfnQP_Wv_dos
XfnQP_Wc_ESeeKfnQP_Wc_E
XfnQP_WcSeeKfnQP_Wc
XfnQP_Wc_dosSeeKfnQP_Wc_dos
DrudeWXd(Units: Energy, Format: Complex) Verbosity:-V resp- Meaning: Drude plasmon energy and inverse lifetime.
- Usage:
EhEngyXd(Units: Energy, Format: Range)Default: (-1,-1) Verbosity:-V resp- Meaning: Electron-hole energy range
- Usage: An alternative way to BndsRnXd to restrict transitions, but will not reduce memory
X_all_q_ROLEs(Units: , Format: )- Meaning: CPUs roles (q,k,c,v)
- Usage: MPI-c,v best memory distribution. MPI-k efficient, some memory replication. MPI-q may lead to load unbalance.
Optics/BSE in eh-space
BEnRange(Units: Energy, Format: Range)- Meaning: Energy range for computing the macroscopic dielectric function using BSE
- Usage:
BDmRange(Units: Energy, Format: Range)Default: (0.10000 , 0.10000)- Meaning: Broadening for the BSE calculation
- Usage: Lorentzian broadening changes linearly between the two values
BEnSteps(Units: None, Format: Integer)- Meaning: Number of evenly spaced energy points in spectrum
- Usage:
BSEBands(Units: None, Format: Range)- Meaning: Bands range: Specifies the band states from which the electron-hole basis of the BSE kernel is constructed
- Usage: Choose few bands close to the Fermi level. Size scales rapidly with the number of bands: BSE is rewritten as an eigenvalue problem for the 2 particle Hamiltonian: size of matrix [Nv x Nc x NKBZ ] x [Nv x Nc x NKBZ]
KfnQPdb(Units: None, Format: String) Verbosity:-V qp- Meaning: Database for QP corrections.
- Usage: From a previous GW calculation. An interpolation of the missing QP-values on the basis of the QP-database is available See also:
KfnQP_E - Format: "<what> < <path>/ndb.QP" with <what> = E,W,Z for QP energy correction, QP width and renormalization factor. For example, from a previous GW calculation use KfnQPdb= "E < ./SAVE/ndb.QP"
KfnQP_E(Units: eV/None/None, Format: Scissors) Verbosity:-V qp- Meaning: QP corrections using a scissor operator and stretching coefficients for the conduction/valence bandwidths.
- Usage: Insert values from a previous GW calculation or experiment. This is a lighter way to include GW corrections than KfnQP_E. Format is: scissor See also:
KfnQPdb
Bethe Salpeter Kernel
BSENGexx(Units: RL/Energy, Format: Integer/Real)- Meaning: G-components to be summed in the Exchange part of the BSE kernel, which takes into account the Local-field effects
- Usage: Small values increase speed. Convergence tests are required.
BSENGBlk(Units: RL/Energy, Format: Integer/Real)- Meaning: Number of RL-components of the Screened Coulomb Potential matrix W(G,G'), to be included in the sum of the e-h attractive Kernel
- Usage: Try using the diagonal terms only first (BSresKmod BScplKmod variables); use a smaller number than the dimension of the Screened interaction matrix
BSHayTrs(Units: None, Format: Real)Default: -0.02- Meaning: Threshold for accuracy of the iterative Haydock process
- Usage: Negative sign: average difference, over the energy range, of two consecutive approximations to the spectrum. Positive sign: maximum difference, over the energy range, of two consecutive approximations to the spectrum
Non-linear
NLtime(Units: fs, Format: Real)Default: -1- Meaning: Simulaiton length
- Usage: If negative automatically set the right value for SHG
- This value is invertionaly proportional to NLDamping
NLCorrelation(Units: , Format: String)Default: IPA (independent particle approximation)- Meaning: Correlation level
- Usage:
HARRLvcs(Units: , Format: Integer)Default: from ns.ndb1- Meaning: Number of G-vectors in the Hartree
- Usage:
EXXRLvcs(Units: , Format: Integer)Default: from Collisions- Meaning: Number of G-vectors in the Screened Exchange
- Usage:
ExtF_Dir(Units: None, Format: Vector)Default: (1,0,0)- Meaning: Direction of the external field
- Usage: