25
28
\item \emph{Electric field and dipole correction for slab calculations}
27
30
Older versions applied an incorrect dipole correction when also using an
28
electric field (old behaviour may be recovered by forcing \texttt{SlabDipoleCorrection .false.}).
31
electric field (old behaviour may be recovered by forcing \code{SlabDipoleCorrection .false.}).
30
Older versions over-estimated the energy correction from the dipole correction by a
33
Older versions over-estimated the energy contribution from the dipole correction by a
31
34
factor of 2 (old behaviour cannot be recovered).
36
\item\emph{Changes in the default precision from grid magnitudes}
38
The grid functions (charge densities, potentials, etc) were in
39
single precision by default in the 3.X versions, but are in double
40
precision by default for post-3.X versions. The \code{phi} array that
41
holds the values of the basis orbitals on the real-space grid is kept in
42
single precision. Please take this into account if you compare the
43
results with those of siesta-3.X runs. See the manual in both
44
versions for more information.
46
\item\emph{Changes in the geometry used for the analysis of the
49
The update of the atomic coordinates after computing the forces on
50
the atoms no longer affects the analysis of the electronic
51
structure at the end of a run. To do the analysis using the actual
52
coordinates used in the computation of the electronic structure, the
53
coordinates are now reset to those values in \code{siesta\_analysis}. The old
54
behavior can be recovered by using by setting the backward
55
compatibility switch \opt{compat-pre-v4-dynamics} to 'true'.
57
This change will slightly alter the output of most calculations and the
58
detailed results of any post-processing. Keep this in mind if you
59
need to maintain coherency within a project.
62
\item\emph{Changes in the handling of the density-matrix and
63
hamiltonian after convergence of the scf cycle}
65
To guarantee that the DM and H used in the electronic-structure
66
analysis correspond to the latest electronic structure computed,
67
there should be no mixing at the end of the scf loop. This
68
behavior, controlled by the value of \opt{scf-mix-after-convergence}, is
69
now the default. The use of DM\_out instead of DM\_mixed provides
70
also a truly variational total-energy and more consistent
71
forces. If the scf cycle is carried out mixing H, H(DM\_in) is
72
recovered by re-setting H to Hold in \code{siesta\_forces} at the end of
75
Also, to preserve the original H for analysis it is necessary to
76
avoid overwriting it in \code{post\_scf\_work}. This is now done by
77
default, except if the option \opt{scf-recompute-h-after-scf} is
80
There is another switch for pre-v4.0 compatibility in this
81
context: If \opt{compat-pre-v4-dm-h} is 'true', the
82
\opt{scf-mix-after-convergence} and \opt{scf-recompute-h-after-scf}
83
options are activated too.
85
This change will slightly alter the output of most calculations
86
and the detailed results of any post-processing. Keep this in mind
87
if you need to maintain coherency within a project.
added
removed
Docs/compatibility.tex
