[Pw_forum] Documentation on atomic velocities

[luisen]

Dear Ilya,

Thanks for your replies.

Just to confirm I understood correctly, are the atomic velocities saved in 
the temporary files so that restarting with restart_mode='restart' 
will resume the dynamics continuously?
And concerning my original questions, can the velocities be written to 
the ordinary output file using some undocumented option?

Many thanks for the support to users. Time is gold and you give it 
away. I really appreciate it.

Luis E. Gonzalez
Universidad de Valladolid.


On Tue, 21 Mar 2017, Ilya Ryabinkin wrote:

> By the way, to resume an interrupted MD run you DON'T need
> ATOMIC_VELOCITIES. They are useful only if you'd like to start (for
> whatever reason) with a setup with non-zero *initial* velocities (like
> I needed once).
>
> The right way is to use restart_mode = 'restart' with all temporary
> files available from the previous (interrupted) step.
>
> The reason is simple: to restart Verlrt algorithm you need more than
> initial velocities.
>
>
> --
> I.
>
> On Tue, Mar 21, 2017 at 9:13 PM, Ilya Ryabinkin <igryabinkin at gmail.com> wrote:
>> Dear Luis:
>> Indeed, there is an undocumented option: pw.x reads the card
>> ATOMIC_VELOCITIES once the option
>>
>> &IONS
>>    ion_velocities = 'from_input',
>> /
>>
>> is provided.
>>
>> They are given in Ry/bohr, no other units are accepted.
>>
>> --
>> Ilya
>>
>>
>> On Tue, Mar 21, 2017 at 6:53 PM, luisen <luisen at metodos.fam.cie.uva.es> wrote:
>>>
>>> Sorry for not being precise enough. I am using the MD calculation in pw.x
>>>
>>> Luis E. Gonzalez
>>> Univ. de Valladolid
>>> Spain.
>>>
>>>
>>> On Tue, 21 Mar 2017, LEUNG Clarence wrote:
>>>
>>>> Dear Luis,
>>>> Do you use the cpmd of cp.x or the md in pw.x?
>>>>
>>>> Thanks.
>>>>
>>>> Clarence
>>>> City university of Hong Kong
>>>>
>>>>
>>>> -------- Original message --------
>>>> From: luisen <luisen at metodos.fam.cie.uva.es>
>>>> Date: 21/03/2017 20:42 (GMT+08:00)
>>>> To: pw_forum at pwscf.org
>>>> Subject: [Pw_forum] Documentation on atomic velocities
>>>>
>>>>
>>>> Dear developers,
>>>>
>>>> I am interested in using atomic velocities generated in previous MD runs
>>>> in order to continue the MD calculation after a finished job (finished
>>>> either gracefully or abruptly).
>>>> I made a small program to compute the velocities from the positions and
>>>> modified slightly the code in order to read the velocities from a file
>>>> upon initialization if this file was present. It is however an ad-hoc
>>>> solution.
>>>>
>>>> I have seen in the last distribution (6.1) that there is the possibility
>>>> of using a section in the input file named ATOMIC_VELOCITIES.
>>>>
>>>> My question is if there is any documentation for its use, and if there
>>>> is any flag so as to output the velocities (in addition to positions) to
>>>> the output file, for later use as input in a subsequent run.
>>>>
>>>> Yours,
>>>>
>>>>          Luis Enrique Gonzalez
>>>>          Universidad de Valladolid
>>>> _______________________________________________
>>>> Pw_forum mailing list
>>>> Pw_forum at pwscf.org
>>>> http://pwscf.org/mailman/listinfo/pw_forum
>>>>
>>>
>>> _______________________________________________
>>> Pw_forum mailing list
>>> Pw_forum at pwscf.org
>>> http://pwscf.org/mailman/listinfo/pw_forum
>>
>>
>>
>> --
>> *******************************************************
>>                     Ilya Ryabinkin
>>                  Postdoctoral Scholar
>>           Physical and Environmental Sciences
>>            University of Toronto Scarborough
>>   http://www.utsc.utoronto.ca/~aizmaylov/Members.html
>> *******************************************************
>
>
>
> -- 
> *******************************************************
>                    Ilya Ryabinkin
>                 Postdoctoral Scholar
>          Physical and Environmental Sciences
>           University of Toronto Scarborough
>  http://www.utsc.utoronto.ca/~aizmaylov/Members.html
> *******************************************************
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>