[Pw_forum] How dose the restart work?

[Gabriele Sclauzero]

Dear Pieremanuele

Pieremanuele Canepa wrote:
> Dear all ,
> my question for today is: How does the restart work in PWscf when in a
> vc-relax job?
> Once activated the restart, as restart_mode = 'restart', in the
> control section, will PWscf read either the last wave function and the
> last geometry or just the last density?

I don't know much about specificity of vc-relax, but in general when restarting pwscf the 
atomic configuration will be read from file (see in prefix.save/data_file.xml) as well as 
the wavefuntions and charge density from the last iteration of the previous run (if the 
files are availalable).
If using wfc_extrapolation and pot_extrapolation, perhaps also older wfcs and older rho 
from previous run will be read from file.

> Would it possible to restart a job from the previous geometry instead
> of the the wave function? 

You can avoid the wavefunction to be read specifying startingwfc='atomic' (or simply 
removing/renaming the file). The code will use a superposition of atomic wavefunctions to 
build the initial guess (as for the first iteration of a run "from_scratch"), but this 
will require more time to perform the first iteration of the scf cycle.

> Can I do both of them?

Both what?


HTH

GS

> 
> All the best, Piero
> --
> Pieremanuele Canepa
> Room 104
> School of Physical Sciences, Ingram Building,
> University of Kent, Canterbury, Kent,
> CT2 7NH
> United Kingdom
> -----------------------------------------------------------
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
> 

-- 


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