lp:~albertog/siesta/trunk-edens

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

704. By Alberto Garcia on 2018-06-19

Use ebs_density internal routine in dhscf

703. By Alberto Garcia on 2018-06-18

Carry out Ebs-density calculation in dhscf

If the fdf incantation

  save-ebs-density T

is used, the program will write a file SystemLabel.EBS (and possibly
other netCDF files) containing the band-structure energy density (just
the density weighted by the eigenvalues of each wavefunction, and
computed by using the energy-density matrix instead of the
density-matrix).

The calculations are carried out in dhscf, as it is the natural place,
even though it is getting too big and its functionality should be
streamlined.

702. By Alberto Garcia on 2018-06-08

Fixes for compilation (still does not do much)

701. By Alberto Garcia on 2018-06-07

Initial survey for Ebs_dens calculation

700. By Alberto Garcia on 2018-06-07

Allow the option of not separating SR and SO in SOC

In a SO calculation, the use of the fdf line:

    soc-split-sr-so F

will prevent the program from splitting the lj NL contributions to the
Hamiltonian and energies into SR (ion) and SO parts. The splitting
procedure for the 'offsite' flavor of SOC can be ill-defined. For the
'onsite' flavor of SOC this problem does not appear, but the option is
also available for generality.

When the SO contribution is not split, the Enl energy is tagged
'Enl(+so)' and the Eso energy is tagged 'Eso(nil)' in the output file.

The CML file is not thus changed (but there is a new parameter
'Split-SR-SO').

Note that this is only a cosmetic change affecting the reporting of
some components of the energy. Internally, the change affects the
H_vkb and H_so components of the Hamiltonian, which are currently only
used to compute the Enl and Eso energies.

699. By Alberto Garcia on 2018-05-30

Merged 4.1, r920-921: YAML infrastructure; TS tests fixes

698. By Nick Papior on 2018-05-27

Restructured k-point sampling

  Now k-points may be user-supplied for further finetuning.
  This may be extremely useful in various situations:
  1) Detailed analysis on PDOS for spin-orbit calculations
  2) Faster convergence with fewer k-points
  3) Improved convergence for bias calculations by adding zoom-in
     k-points in the bias-window.
  4) And others.

  Now the kpoints are associated with a type:
    kpoint_t
  this will allow one to read k-points using kgrid.File flag.
  The precedence is much like the older implementation:

  1) kgrid.MonkhorstPack
  2) kgrid.Cutoff
  3) kgrid.File

  in that order. If none are specified, then the Gamma-point will be used.

  This meant that several places the code needed updating.
  - Projected DOS k-points
  - TranSiesta k-points
  - SCF k-points

  Now all references to the kpoint_grid module is using kpoint_scf_m module
  which uses the type.
  This not only reduces lots of duplicated code, but it also allows
  one implementation to benefit across all code segments.

  Also added a test that reads in the k-points from a file (si_coop_kp_file).

697. By Alberto Garcia on 2018-05-19

Merged 4.1, r919: io_close placement

696. By Alberto Garcia on 2018-05-15

Merged 4.1, r910-918

695. By Nick Papior on 2018-05-07

Merged r905-909, small fixes here and there, improved print-out for ts