[Q-e-developers] a specific question on the whole structure of pwscf

[Paolo Giannozzi]

On Sat, 2011-11-19 at 21:54 -0600, Jun Liu wrote:

> QE introduces itself that it can use norm-preserving, ultrasoft and
> PAW methods to calculation total energy and bandstructure, I am
> wondering whether these three different modes are independently
> implemented, or they are inter-connect with each other.

the ultrasoft case is an extension of the norm-conserving 
case (i.e. US = NC + additional pieces). I don't know well 
the PAW case, but it should be like the US case, with a few 
terms treated in a different way, and a few PAW-specific terms

> In both cases, which part should I focus on if I am only 
> interested in PAW method?

it depends on what you need to implement. The PAW-specific
code is mostly in paw_onecenter.f90; other paw-specific stuff
is in paw_init.f90 and paw_symmetry.f90

PG
-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy