CASINO manual - Theory of Condensed Matter

PERMIT DEN SYMM (Logical) If this flag is set to T then he symmetry of the SCF charge
density (in mpc.data) will be imposed on the QMC charge-density data used in the MPC
interaction (with the justification that imposing an exact condition on the charge density can’t
hurt) and also when writing the QMC density to expval.data. It is possible however that
DMC will break the symmetry of the SCF calculation; in this case the user should turn off
permit den symm.
PLOT BACKFLOW (Block) This block allows a plot of the backflow transformation to be made
(see Sec. 23) just after VMC equilibration. The block should contain 2 lines, plus an optional
line for plotting the Φ term: (1) ‘0’ or ‘1’ to (de-)activate this facility; (2) ‘kspin’, ‘knumber’
and ‘zposition’; (3) value of fixed electron–nucleus distance r iI . This will produce various files:
bfconfig.dat (reference config), bfconfigx.dat (associated quasi-particle config), bfions.dat
(coordinates of nuclei for which backflow terms exist), bfeta 〈s〉.dat (η vs. r ij for each spinpair
type 〈s〉), bfmu 〈s〉 〈set〉.dat (µ vs r i for each spin type 〈s〉 in each set 〈set〉), bfphi.dat
[contribution of Φ to the 3D backflow displacement on electron j vs. 2D projection of r jI on the
plane defined by electron j, ion I in set 〈set〉, and electron i at distance r iI from the nucleus with
spin such that 〈s〉 is the spin-pair type of (i, j)], and bffield.dat (3D backflow displacement
on electron (kspin,knumber) vs. its 2D position on the plane z = zposition).
PLOT EXPVAL (Block) This block allows the plot expval utility to be told about the geometrical
region over which a particular expectation value is to be plotted (which can be a line AB / plane
AB-AC / volume AB-AC-AD). The structure of the block is as follows:
Line 1: dimensionality d (1, 2 or 3);
Line 2: No. of points along the d directions;
Line 3: x, y and z coordinates of point A;
Line 4: x, y and z coordinates of point B;
Line 5: x, y and z coordinates of point C (if required);
Line 6: x, y and z coordinates of point D (if required).
The data will be plotted in a format suitable for xmgrace in the file lineplot.data or in
a format suitable for gnuplot in 2Dplot.dat or 3Dplot.dat which can be quickly visualized
using the plot 2D utility. This block is ignored by casino itself.
POPSTATS (Logical) If popstats is T then the variance of the local energies sampled in DMC will
be evaluated.
POSTFIT VMC (Logical) If postfit vmc is set to T then a VMC calculation will be performed
after the final optimization phase when runtype=‘vmc opt’ or ‘opt vmc’. This should normally
be done.
POSTFIT KEEP CFG (Logical) If postfit keep cfg is set to T then the configurations generated
in the post-fit VMC calculation will be written to config.out.
PRIMITIVE CELL (Block) Sometimes the ‘primitive lattice vectors’ in the xwfn.data file do not
correspond to the true primitive cell. If the actual primitive lattice vectors are needed (e.g., for
accumulation of the charge density), then override values can be supplied in the primitive cell
block.
PRINTGSCREENING (Logical) Before doing a periodic Gaussian calculation, casino prepares
lists of potentially significant (primitive) cells and sites in each such cell which could contain
Gaussians having a nonzero value in a reference primitive cell centred on the origin. Zero is
defined as 10 −gautol . Turning on the printgscreening flag prints out the important information
about this screening.
PSI S (Text) If psi s is ‘none’ then the Slater wave function is set to one; if psi s is ‘slater’ (default)
then an expansion in one or more Slater determinants is used; if psi s is ‘exmol’ then a specially
crafted wave function for excitonic and positronic molecules is used.
QMC DENSITY MPC (Logical) If this flag is set to T then the QMC charge data at the end of
the expval.data file will be used to compute the MPC interaction (see Sec. 19.4.4), rather than
the default SCF density in the mpc.data file. This is likely to be useful in cases such as the
Wigner crystal where the Hartree-Fock charge density is very different to the true charge density
(it is too localized) as opposed to, say, the Fermi fluid where the Hartree-Fock charge density
is exact. Note that using this option is likely to increase the time taken to evaluate the MPC
interaction; in both DFT and QMC cases, the code counts backwards from the end of the list
50