[Pw_forum] What files are needed when pw.x restarts? In what case the *.wfc files are need?

Giuseppe Mattioli giuseppe.mattioli at mlib.ism.cnr.it
Wed Jul 27 11:44:25 CEST 2011


Dear Liu
Only a couple of notes:
The wf_collect key collects indeed the wfc information (at the end of the scf 
cycle, I suppose, but I'm not sure) in order to restart the calculation with 
a different parallel scheme. It is particularly useful when you want to 
perform serial post-processing calculations (i.e., when using the pp.x code) 
after parallel pw.x ones.
However, pw.x tries to restart the calculation by looking for everything 
useful in a number of different possible schemes: for instance, when you 
perform a geometry optimisation and the .wfc* files are badly written or 
corrupted, you may restart from scratch the scf cycle, without affecting the 
bfgs history contained into the .bfgs binary file.
HTH
Giuseppe

On Wednesday 27 July 2011 09:41:59 Seyed Javad Hashemifar wrote:
> I guess if you set wf_collect=.true. the collection of *.wfc files is done
> after reaching self consistency not after each iteration. Probably, the
> evc.dat files are the collected *.wfc files (I am not sure). You may check
> their sizes, if they are very big files, then most likely they are the
> collected wave functions files.
> SJ Hashemifar
> ======================================
> Seyed Javad  Hashemifar
> Department of Physics, Isfahan University of Technology
> 84156-83111 Isfahan, Iran
> Tel: +98 311 391 2375     Fax:+98 311 3912376
> Email:  hashemifar at cc.iut.ac.ir
> Group Homepage:      http://cmsgroup.iut.ac.ir
> Personal Homepage:  http://hashemifar.iut.ac.ir
> ---------------------------------------------------------------------------
>
> On Wed, Jul 27, 2011 at 9:28 AM, gbliu <goodluck_1982 at 163.com> wrote:
> > **
> > Hi Hashemifar,
> >
> > I just guess and expect that  wf_collect=.true. will collect *.wfc files
> > from wfcdir to outdir after each iteration or at the end of run. Is this
> > right? But it seems not.  As mentioned in last mail, *.wfc files don't
> > appear in ourdir  anytime.
> >
> > Best wishes
> > Liu Guibin
> >
> >
> > 于 2011/7/27 14:13, Seyed Javad Hashemifar 写道:
> >
> > Hello
> > As far as I know, since pw.x uses iterative diagonalization method, in
> > addition to the charge density files, it needs to wave function files
> > also to restart a calculation. If there is no wave function file, then
> > you can set some atomic wave functions in namelist ELECTRONS:
> >
> >  http://www.quantum-espresso.org/input-syntax/INPUT_PW.html#id2468572
> >
> >  to restart your job, but in this case it will not be restarted exactly
> > from the position of interruption.
> >
> >  If you set wf_collect=.true. then most likely you able to restart or
> > continue your parallel job with a different parallel setting (for example
> > on different number of processors).
> > SJ Hashemifar
> > ======================================
> > Seyed Javad  Hashemifar
> > Department of Physics, Isfahan University of Technology
> > 84156-83111 Isfahan, Iran
> > Tel: +98 311 391 2375     Fax:+98 311 3912376
> > Email:  hashemifar at cc.iut.ac.ir
> > Group Homepage:      http://cmsgroup.iut.ac.ir
> > Personal Homepage:  http://hashemifar.iut.ac.ir
> > -------------------------------------------------------------------------
> >--
> >
> > On Wed, Jul 27, 2011 at 7:04 AM, gbliu <goodluck_1982 at 163.com> wrote:
> >>  Dear all,
> >>
> >>      Now I'm wondering what files are needed when pw.x restarts from a
> >> interrupted run. Is a single file outdir/prefix.save/charge-density.dat
> >> enough? Are *.wfc files needed?
> >>      What's the function of the parameter "wf_collect = .true." ?  I set
> >> outdir='./work' , wfcdir='/tmp', restart_mode='restart'. Then *.wfc
> >> files exist  in /tmp.
> >>      I note that, when I set *wf_collect = .false.*, there are files
> >> under ./work  (case 1):
> >>         xxx.UPF
> >>         charge-density.dat
> >>         data-file.xml
> >>        K00001/eigenval.xml
> >>        K.....
> >>    and when I set *wf_collect = .true.*, there are files under ./work
> >> (case 2):
> >>         xxx.UPF
> >>         charge-density.dat
> >>         data-file.xml
> >>         gvectors.dat
> >>        K00001/eigenval.xml
> >>        K00001/evc.dat
> >>        K00001/gkvectors.dat
> >>       K .....
> >> What's the use of gvectors.dat files?
> >>  In case 1, because wf_collect=.false., *.wfc files keep existing in
> >> /tmp,  when pw.x restarts, it can read wfc from /tmp.
> >> However, in case 2, *.wfc files are removed from /tmp after run and
> >> don't appear in ./work either, therefore when pw.x restarts, it cannot
> >> read wfc from anywhere. Now that wf_collect=.true., how are wfc files
> >> are collected and to where?
> >> BTW:  in both cases, pw.x can read the existing charge density, is this
> >> enough for a restarted run even if  no wfc files found?
> >>
> >> Best wishes
> >> Liu Gui-Bin
> >> Dept. of Physics
> >> The Universit of Hong Kong
> >>
> >> _______________________________________________
> >> Pw_forum mailing list
> >> Pw_forum at pwscf.org
> >> http://www.democritos.it/mailman/listinfo/pw_forum
> >
> > _______________________________________________
> > Pw_forum mailing
> > listPw_forum at pwscf.orghttp://www.democritos.it/mailman/listinfo/pw_forum
> >
> >
> >
> > _______________________________________________
> > Pw_forum mailing list
> > Pw_forum at pwscf.org
> > http://www.democritos.it/mailman/listinfo/pw_forum



-- 
********************************************************
- Article premier - Les hommes naissent et demeurent
libres et ègaux en droits. Les distinctions sociales
ne peuvent être fondèes que sur l'utilitè commune
- Article 2 - Le but de toute association politique
est la conservation des droits naturels et 
imprescriptibles de l'homme. Ces droits sont la libertè,
la propriètè, la sùretè et la rèsistance à l'oppression.
********************************************************

   Giuseppe Mattioli                            
   CNR - ISTITUTO DI STRUTTURA DELLA MATERIA   
   v. Salaria Km 29,300 - C.P. 10                
   I 00015 - Monterotondo Stazione (RM)          
   Tel + 39 06 90672836 - Fax +39 06 90672316    
   E-mail: <giuseppe.mattioli at ism.cnr.it>



More information about the users mailing list