Tutorial tarball added, convolution of diffuse scattering maps becomes automatically from disca.x and ZG.x without extra postprocessing (i.e. no pp_disca.x is needed). Input flags are described in "doc/input_variables.pdf". JDOS example is added in the tutorial tarball.
Module "spin_orb" merged into "noncollin_module": there is little logic in
having two small modules ("upf_spinorb" is the other one) for a handful of
spin-orbit related variables. Morever "noncollin_module" and "spin_orb" are
almost everywhere USEd together.
BEWARE: may break external codes using QE routines.
The list of all general-purpose modules is defined once and for all in make.inc
and used everywhere. Can be further improved. IMPORTANT: re-run configure!
ZG folder is renamed from "ZG_displacement" to "ZG" (looks nicer)
tar file containing the example of JDOS is replaced by the folder "example"
"doc" folder is added containing a tutorial and a file describing the input variables
README is also updated accordingly (minor changes)
From a physic point of view, (G+q) should be in unit of 1/Bohr.
It is much easier to deal with these units, especially when adding higher-order multipole.
I therefore change (G+q) from no unit to 1/Bohr unit by multiplying by 2pi/alat
Note that this is purely cosmetic for all cases except when quadrupole is use.
In such case there is a correct additional 2pi/alat factor.
In the case of rgd_blk_epw and rgd_blk_epw_fine the Ewald exponential should be adimentional.
However they are not at present because of this re-definition "qeq = qeq * twopi / alat".
For now I kept this inconsitency but we should make it adimensional in the future or (better) remove the
Ewald filter.
1) Add degeneracy treatment to dmewan2bloch to mirror the one done in vmewan2bloch
2) change the vme input from logical to CHARACTER with value 'dipole' and 'wannier'.
3) Upgrade version number of EPW to v5.4b
4) Add a warning to user when using system_2d input