[xcrysden] Fermi surface and pwscf

Tone Kokalj xcrysden@democritos.it
Fri, 18 Aug 2006 18:01:53 +0200


On Mon, 2006-08-14 at 12:06 -0700, Jesse Noffsinger wrote:
> My two questions are: 
> 1) Is the grid of points in cartesian coordinates or along the
> primitive vectors? (your example is a sc structure, so it's hard to
> tell)

The grid of points is "along" the primitive reciprocal lattice vectors.
In other words: the grid of points is a uniform grid-of-points contained
in the reciprocal unit cell.

> 2) When I tell pwscf to calculate a 20x20x20 grid I get 4000 pts (so
> 1/2 of what I should).  I know this isn't your area, but maybe you've
> had experience with this -- where do I get the rest of the points? 

Because of the time-reversal symmetry. If you have a simmetric
structure  
the PWscf (or any other ab inito program) will generate even lesser
number of k-points ...

Regards, Tone