[Octopus-users] td run: Only 384 files out of 64 could be read.
Xavier Andrade
xavier at tddft.org
Wed Sep 15 21:46:06 WEST 2010
Hi Emily,
The problem is that the td_occup option does not work with state
parallelization (since the states might be in another node, it is tricky
to calculate the projection in this case). In the development version I
think there is a more precise error message.
As a workaround, you can run the calculation in serial or use domain
parallelization.
Cheers,
Xavier
On Wed, 15 Sep 2010, Emily wrote:
> I just discovered that this problem goes away if I remove td_occup from
> the line
> TDOutput = laser + energy + multipoles + td_occup
> in the input file.
>
> There is very little documentation of this feature...what I know about
> it is:
> td_occup (256): If set, outputs the projections of the time-dependent
> Kohn-Sham wave-functions onto the static (zero time) wave-functions to
> the file td.general/projections.XXX.
>
> Can anybody hint what I am doing wrong?
>
> (I am trying to do something similar to Quijada, Muino, Borisov, Alonso
> & Echenique, New Journal of Physics 12 (2010) 053023, in which they
> project the time dependent KS states onto the ground state.)
>
> Thanks,
> Emily
>
> ----- Original message -----
> From: "Emily" <listemily at eml.cc>
> To: "octopus users group" <octopus-users at tddft.org>
> Date: Wed, 15 Sep 2010 12:12:28 -0400
> Subject: td run: Only 384 files out of 64 could be read.
>
> Can anybody suggest to me why my time dependent run can't find my ground
> state restart files? If I run another gs run, it finds the ground state
> files and converges in only 4 cycles. When I look in restart/gs/ I see
> the 64 obf files, along with a few other files. But when I do a time
> dependent run I get the complaint that "only" 384 files out of 64 could
> be read. (64*6 = 384, is this significant?) Snips of the relevant
> output and inp file below.
>
> Thanks,
> Emily
>
> Output:
>
> **********************************************************************
>
> Info: Loading restart information.
> Could not load any previous restart information.
> **********************************************************************
>
>
> ** Warning:
> ** Could not load restart/td: Starting from scratch
>
> Info: Loading restart information.
> Info: Loading restart information.
>
> ******************** Loading restart information. ********************
> Only 384 files out of 64 could be read.
> **********************************************************************
>
>
> **************************** FATAL ERROR *****************************
> *** Fatal Error (description follows)
> *--------------------------------------------------------------------
> * From node = 0
> *--------------------------------------------------------------------
> * Could not load restart/gs
> **********************************************************************
>
> *** Fatal Error (description follows)
> *--------------------------------------------------------------------
> * From node = 4
> *--------------------------------------------------------------------
> * Could not load restart/gs
> *** Fatal Error (description follows)
> *--------------------------------------------------------------------
>
> ___________________________________________________
> parts of my inp file (leaving out parts that define the system and the
> box, as I don't think they would be relevant.):
>
> temp = 1/1000
> Smearing = temp
> SmearingFunction = fermi_dirac
> # CalculationMode = gs
> CalculationMode = td
> Units=atomic
> ExtraStates=14
>
> ### time dependent section
> amplitude = .0001
> omega = 158/1000
>
> %TDFunctions
> "constant" | tdf_cw | amplitude
> %
>
> %TDExternalFields
> electric_field | 0 | 1 | 0 | omega | "constant"
> %
>
> tmax = 1000 #
> TDEvolutionMethod = aetrs
> TDTimeStep = 0.1 #
> TDMaximumIter = tmax/TDTimeStep
> OutputEvery = 10
> TDOutput = laser + energy + multipoles + td_occup
>
> _______________________________________________
> Octopus-users mailing list
> Octopus-users at tddft.org
> http://www.tddft.org/mailman/listinfo/octopus-users
>
More information about the Octopus-users
mailing list