[Pw_forum] Phonons at generic q-point

[Dal Corso Andrea]

This bug has been fixed in the cvs version but not yet in the 4.0.3
version.
Please add the line:

twfcollect=.false.

at line 285 of PH/phonon.f90

284         CALL electrons()
285         !
286         CALL close_files()

Between CALL electrons and CALL close_files.

And change also line 20 of phonon.f90 with:

   USE control_flags,   ONLY : gamma_only,  twfcollect


Then recompile and try again. 

The alternative is not to use wf_collect=.true. in the pw calculation. 

Andrea



On Tue, 2008-11-04 at 17:31 +0000, Benedek, Nicole A wrote:
> Dear All,
> 
> I have been trying to run a parallel phonon calculation at generic
> q-point for a SrTiO3 multilayer (16 atoms) using espresso 4.0.3. I run
> an initial scf calculation with wf_collect=true and then a ph.x
> calculation with lnscf=true. The calculation always crashes with the
> following error (full input pasted below):
> 
> ! ... rest of modes
> 
>      Representation    47      1 modes - To be done
> 
>      Representation    48      1 modes - To be done
> 
>      from openfilq : error #         1
>      file sto_mag.wfc not found
>      stopping ...
> 
> The wfc files are present and written to the same directory from which
> the calculation is running and this is not a 'recovered' calculation.
> 
> So far I have tried the following tests:
> 
> * The examples for Si at http://people.sissa.it/~degironc/QE-Tutorial/
> (all run successfully);
> * Phonon dispersion and q=0 calculations for bulk SrTiO3 (all run
> successfully);
> * Phonon calculation at M for bulk SrTiO3, with lnscf=true (run
> successfully);
> * Phonon calculation at q=0 for SrTiO3 multilayer (run successfully).
> 
> I also noticed the message 'Found additional translation' printed at
> the top of the output. I had a look through the source code
> (sgam_at.f90 and sgam_at_mag.f90), so I know this message is printed
> if the identity is found to have fractional translations. I found one
> previous post about this message in the archive
> (http://tinyurl.com/5k926j) but it seemed to concern how parallel jobs
> were re-started before the wf_collect keyword existed (I run all my
> calculations with wf_collect=true and my phonon calculation uses the
> same number of processors as the initial scf run). Is this message the
> source of my problems or is there something else going on? Any hints
> would be appreciated.
> 
> Cheers,
> 
> Nicole
> 
> Nicole Benedek
> Research Associate
> Department of Materials
> Imperial College London
> 
> ! INPUT FILE FOR PHONON CALC
> 
> Phonons of SrTiO3 multilayer at X
>  &inputph
>   tr2_ph=1.0d-14,
>   amass(1)=87.62,
>   amass(2)=47.88,
>   amass(3)=15.9994,
>   prefix='sto_mag',
>   outdir='./'
>   fildyn='sto_mag.dynX',
>   lnscf=.true.
>  /
>  0.0 0.5 0.0 ! X point in BZ for orthorhombic primitive systems
> 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
-- 
Andrea Dal Corso                    Tel. 0039-040-3787428
SISSA, Via Beirut 2/4               Fax. 0039-040-3787528
34014 Trieste (Italy)               e-mail: dalcorso at sissa.it