[Pw_forum] indexing of G and k+G vectors.
Stefano de Gironcoli pw_forum@pwscf.org
Mon, 14 Apr 2003 17:38:41 +0200

    * Previous message: [Pw_forum] (no subject)
    * Next message: [Pw_forum] Installation Error
    * Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Dear user,

the list of G vectors g(3,ngm) contains all plane wave with kinetic energy
below ecutrho in order of increasing length.

The index igk(npw) created by calling
gk_sort (xk (1, ik), ngm, g, ecutwfc / tpiba2, npw, igk, g2kin)
for each kpoint when needed (or reading from the file  'prefix'.igk, 
unit=iunigk)
contains the information in order to go from the ordering of increasing  
k+G
length (the one used for eigenvectors) to the ordering od increasing G 
length
(used for charge density, potential etc).

That is:  k+G(igk(i)) is the vector that comes in the i-th place in the 
list of k+G
ordered by their length

g2kin(i) is the corresponding kinetic energy in unit of tpiba2 
[=(2pi/alat)^2]


best regards,

   Stefano de Gironcoli

Raghani Pushpa wrote:

>Dear user,
>There is variable in the code called 'npwx'. I thought it is the number of 
>plane waves i.e. the no of G vectors. However, there is a difference of 
>factor of 8 between them, i.e. no of G vectors ~ 8*npwx. What does this 
>variable physically mean? 
>
>How can I get the coefficients C(k+G) from the code and how will I get the 
>correspondence between C(k+G) and G vectors i.e. which C(k+G) corresponds 
>to which G vector?
>
>Thanx,
>Pushpa
>
>  
>