[Pw_forum] Changing nr3

[Paolo Giannozzi]

On Saturday 02 March 2013 11:31, Varadharajan Srinivasan wrote:

> For instance, if I try the following input file :
> [...]  &system
>     ibrav=  2, celldm(1) = 10.35, nat=  2, ntyp= 1,
>     ecutwfc= 20.0,
>     ecutrho=80.0,
>     nr3 = 28 [...]
> I get the following in the output     (24 instead of the 28 that I ask for)
>
>  Dense  grid:     3383 G-vectors     FFT dimensions: (  24,  24,  24)

you have to set all three FFT dimensions, not just one.
See Modules/griddim.f90, subroutine realspace_grids_init. 
By the way, in your case the system has three crystal axis 
of the same length. You shouldn't set them to different FFT 
dimensions. While the main code should still work, there
will be at least some cases in which symmetrization will not
work if the FFT grid breaks the symmetry of the crystal.

P.
---
Paolo Giannozzi, IOM-Democritos and DCFA, Univ. Udine