[Pw_forum] can't read wavefunction and coords

Gabriele Sclauzero sclauzer at sissa.it
Mon Apr 27 16:37:20 CEST 2009

Jonas Baltrusaitis wrote:
> Gabriele,
> thanks for your response. I reran input without the SCF step. I don't thin it's reading
> coords right. The source of confusion is that I am inputting and outputing coords in
> angstroms in optimization but it gets transferred into alat in normal mode calcs.

I think reading and writing coordinates is treated properly in pwscf (i.e. it is 
internally consistent). You can specify input coordinates in different formats, alat 
(Cartesian axes, celldm(1) units), crystal,...
At the beginning of the output the read coordinates are written in alat units (correctly 
transformed, if you used an input with a different format), and in crystal units if you 
use verbosity='high'.
In relax calculations the coordinated are written after each scf step in the format you 
used for the input. So, if you use (angstrom) as input format, you'll get ATOMIC_POSITION 
in angstrom units on output, so that you can copy and paste those positions if you need 
them to build another input.
For me it always worked: if I give input coordinates as

I will get

after each ions move in a relax calculation.
If you restart, the program should not care about the atomic positions you give on input, 
but read those store in the restart files.



> Still, this is what I backcalculate with xcrysden:
> Optimized coords: ATOMIC POSITIONS Si -0.053476533 -0.187541319 +0.057950773 ...
> Normal mode calculation coords: ATOMIC_POSITIONS Si      -0.037813604  -0.132611686
> 0.040977368 H       -0.042235675  -0.102590555   0.268983495 H       -0.221511706
> -0.243919610  -0.030796850 H       -0.032400640   0.077902395  -0.042412224 H
> 0.144511653  -0.248234756  -0.027243038
> below is my latest input incidently, is there a way to tell pwscf to use only angstroms
> in output? 
> It is convenient for me to use angstroms when setting up the initial input
> so I'd like to stay with angstroms
> JOnas


o ------------------------------------------------ o
| Gabriele Sclauzero, PhD Student                  |
| c/o:   SISSA & CNR-INFM Democritos,              |
|        via Beirut 2-4, 34014 Trieste (Italy)     |
| email: sclauzer at sissa.it                         |
| phone: +39 040 3787 511                          |
| skype: gurlonotturno                             |
o ------------------------------------------------ o

More information about the users mailing list