The ABINIT Users Mailing List ( CLOSED )

[verstraete@pcpm.ucl.ac.be]

Hi Jim, 

What Ia think is happening is as follows: abinit reduces any grid it can 
to a (not necessarily simple-cubic) unique grid wqith one shift. The 
lattice vectors for the kpoints are obtained from kptrlatt. For FCC, for 
instance, ngkpt 4 4 4 and nshiftk 4 etc... becomes

shiftk 0.5 0.5 0.5
nshiftk 1
and 
kptrlatt 4 -4 -4   4 -4 4   -4 -4  4
(which is exactly the same kpoints!)

I haven't examined the kpoints you get in v3_t83, but my money's there.

In some wierd cases this isn't possible if the kpoints aren't homogeneous 
enough, but I haven't worried about this too much. All Monkhorst Pack 
grids can be cast into a single kptrlatt + 1 shiftk.

Ciao

Matthieu


On Mon, 24 Mar 2003, James Raynolds wrote:

> Dear abinitioners,
> 
>       If you do a "grep nshiftk *.out" in Test_v3/Refs you will find no
> reference to the variable nshiftk and you will notice that in all runs it
> has been implicitly set to the default nshiftk = 1.  Thus when asking for
> nshiftk = 4 (four shifted grids) as in t83.in you only get one!
> 
>       Am I missing something or is this a bug?
> 
> Sincerely,
> Jim Raynolds
> 

-- 
===================================================================
Matthieu Verstraete               mailto:verstraete@pcpm.ucl.ac.be 
PCPM, Boltzmann, pl. Croix du Sud, 1            tel: 010/ 47 86 81
B-1348 Louvain-la-Neuve  Belgium                fax: 010/ 47 34 52