Question about use of REFTRAJ

[Rajorshi Chattopadhyay]

Hello all,

I am trying to use REFTRAJ in cp2k to do Wannier Localisation of 100 snapshots from a production simulation. I have two questions:

1. I see that a timestep of 0.5fs is set although I do not set this explicitly in the input file. Since I want to run a static calculation on each snapshot, I do not understand the utility of setting a timestep. Does this have any effect on the calculations ? The relevant portion of my input file is as follows: 

&MD

    ENSEMBLE REFTRAJ

    STEPS 100

    &REFTRAJ

      EVAL_ENERGY_FORCES

      TRAJ_FILE_NAME traj_wan.xyz

    &END REFTRAJ

   &END MD

(traj_wan.xyz contains coordinates of 100 snapshots)

2. I see that there is  non convergence of the scf in some of the snapshots when I run them with the above setting. However, when I run 100 separate calculations (1 for each snapshot), there are no convergence issues. I do not understand why this happens.

Thank you in advance for your help.

--

With Regards,

Rajorshi Chattopadhyay,

PhD Researcher in Mineralogy/Crystallography,

Institut für Geologie und Mineralogie,

Universität zu Köln, Germany

[Jürg Hutter]

Hi

1) the timestep is not used in your case. You can ignore it.
2) default is to extrapolate from previous timestep using the ASPC method.
If the frames are not correlated it is better to start with an independent
initial guess.
-->
FORCE_EVAL / DFT / QS / EXTRAPOLATION USE_GUESS
--> this will make use of
FORCE_EVAL / DFT / SCF / SCF_GUESS ???

regards

Juerg Hutter

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Rajorshi Chattopadhyay <rajors...@gmail.com>
Sent: Tuesday, July 26, 2022 3:46 PM
To: cp...@googlegroups.com
Subject: [CP2K:17385] Question about use of REFTRAJ

Hello all,

I am trying to use REFTRAJ in cp2k to do Wannier Localisation of 100 snapshots from a production simulation. I have two questions:

1. I see that a timestep of 0.5fs is set although I do not set this explicitly in the input file. Since I want to run a static calculation on each snapshot, I do not understand the utility of setting a timestep. Does this have any effect on the calculations ? The relevant portion of my input file is as follows:

&MD

ENSEMBLE REFTRAJ

STEPS 100

&REFTRAJ

EVAL_ENERGY_FORCES

TRAJ_FILE_NAME traj_wan.xyz<http://traj_wan.xyz>

&END REFTRAJ

&END MD

(traj_wan.xyz<http://traj_wan.xyz> contains coordinates of 100 snapshots)

2. I see that there is non convergence of the scf in some of the snapshots when I run them with the above setting. However, when I run 100 separate calculations (1 for each snapshot), there are no convergence issues. I do not understand why this happens.


Thank you in advance for your help.

--
With Regards,
Rajorshi Chattopadhyay,
PhD Researcher in Mineralogy/Crystallography,
Institut für Geologie und Mineralogie,
Universität zu Köln, Germany

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CACTJSRfmRjRBD%2B0GPdRfn6yTTEx-FB-g1fKnpcaeNdmpVwT0%2BA%40mail.gmail.com<https://groups.google.com/d/msgid/cp2k/CACTJSRfmRjRBD%2B0GPdRfn6yTTEx-FB-g1fKnpcaeNdmpVwT0%2BA%40mail.gmail.com?utm_medium=email&utm_source=footer>.

[Rajorshi Chattopadhyay]

Hi Dr. Jürg Hutter,

I thank you for your reply. These solved the problems. 

Additionally, I had one more question after REFTRAJ. I see that I have to provide a coordinate file under &COORD in addition to the trajectory file under &REFTRAJ. I suppose the utility of this coordinate file is to allocate space for the atoms. The coordinates for property calculation comes from the trajectory file under &REFTRAJ. So, in principle the number of atoms and their types should be the same in the coordinate file under &COORD and each snapshot in the trajectory file under &REFTRAJ. Am I right in my assessment ?

I thank you for your help and suggestions.

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759F67E3F50E80D90188AAE9F979%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.

[Jürg Hutter]

Yes, the COORD section defines the structure of the system. Number and type of atoms
and their ordering.

regards

JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Rajorshi Chattopadhyay <rajors...@gmail.com>

Sent: Friday, August 5, 2022 1:01 PM
To: cp...@googlegroups.com
Subject: Re: [CP2K:17436] Question about use of REFTRAJ

Hi Dr. Jürg Hutter,

I thank you for your reply. These solved the problems.

Additionally, I had one more question after REFTRAJ. I see that I have to provide a coordinate file under &COORD in addition to the trajectory file under &REFTRAJ. I suppose the utility of this coordinate file is to allocate space for the atoms. The coordinates for property calculation comes from the trajectory file under &REFTRAJ. So, in principle the number of atoms and their types should be the same in the coordinate file under &COORD and each snapshot in the trajectory file under &REFTRAJ. Am I right in my assessment ?

I thank you for your help and suggestions.

On Wed, 27 Jul 2022 at 10:25, Jürg Hutter <hut...@chem.uzh.ch<mailto:hut...@chem.uzh.ch>> wrote:
Hi

1) the timestep is not used in your case. You can ignore it.
2) default is to extrapolate from previous timestep using the ASPC method.
If the frames are not correlated it is better to start with an independent
initial guess.
-->
FORCE_EVAL / DFT / QS / EXTRAPOLATION USE_GUESS
--> this will make use of
FORCE_EVAL / DFT / SCF / SCF_GUESS ???

regards

Juerg Hutter

________________________________________

From: cp...@googlegroups.com<mailto:cp...@googlegroups.com> <cp...@googlegroups.com<mailto:cp...@googlegroups.com>> on behalf of Rajorshi Chattopadhyay <rajors...@gmail.com<mailto:rajors...@gmail.com>>

Sent: Tuesday, July 26, 2022 3:46 PM

To: cp...@googlegroups.com<mailto:cp...@googlegroups.com>

Subject: [CP2K:17385] Question about use of REFTRAJ

Hello all,

I am trying to use REFTRAJ in cp2k to do Wannier Localisation of 100 snapshots from a production simulation. I have two questions:

1. I see that a timestep of 0.5fs is set although I do not set this explicitly in the input file. Since I want to run a static calculation on each snapshot, I do not understand the utility of setting a timestep. Does this have any effect on the calculations ? The relevant portion of my input file is as follows:

&MD

ENSEMBLE REFTRAJ

STEPS 100

&REFTRAJ

EVAL_ENERGY_FORCES

TRAJ_FILE_NAME traj_wan.xyz<http://traj_wan.xyz><http://traj_wan.xyz>

&END REFTRAJ

&END MD

(traj_wan.xyz<http://traj_wan.xyz><http://traj_wan.xyz> contains coordinates of 100 snapshots)

2. I see that there is non convergence of the scf in some of the snapshots when I run them with the above setting. However, when I run 100 separate calculations (1 for each snapshot), there are no convergence issues. I do not understand why this happens.


Thank you in advance for your help.

--
With Regards,
Rajorshi Chattopadhyay,
PhD Researcher in Mineralogy/Crystallography,
Institut für Geologie und Mineralogie,
Universität zu Köln, Germany


--
You received this message because you are subscribed to the Google Groups "cp2k" group.

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com><mailto:cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com>>.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CACTJSRfmRjRBD%2B0GPdRfn6yTTEx-FB-g1fKnpcaeNdmpVwT0%2BA%40mail.gmail.com<https://groups.google.com/d/msgid/cp2k/CACTJSRfmRjRBD%2B0GPdRfn6yTTEx-FB-g1fKnpcaeNdmpVwT0%2BA%40mail.gmail.com?utm_medium=email&utm_source=footer>.

--
You received this message because you are subscribed to the Google Groups "cp2k" group.

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com>.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759F67E3F50E80D90188AAE9F979%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.


--
With Regards,
Rajorshi Chattopadhyay,
PhD Researcher in Mineralogy/Crystallography,
Institut für Geologie und Mineralogie,
Universität zu Köln, Germany

​

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CACTJSRfnqPi0pk8DMuVw%2Bk5ow%2BFvaOX1FQJJpAs-8RROce9oog%40mail.gmail.com<https://groups.google.com/d/msgid/cp2k/CACTJSRfnqPi0pk8DMuVw%2Bk5ow%2BFvaOX1FQJJpAs-8RROce9oog%40mail.gmail.com?utm_medium=email&utm_source=footer>.