[Pw_forum] Silent Calculations
Stefano de Gironcoli
degironc at sissa.it
Fri Aug 8 17:03:53 CEST 2008
there should be no incompatibility between shifted grids and phonon
calculations.
stefano
Subhradip Ghosh wrote:
> Dear Eyvaz,
>
> It would be nice if you explain why phonons and shifted k-mesh are
> incompatible. Hiowever, I have used shifted k-mesh for phonon calculations
> and didn't find anything unusual for the systems I tried so far.
>
> Regards,
>
> Subhradip
>
>
>
>
>> Dear Miguel,
>>
>> It might be I am wrong, but I am not sure phonons and shifted k-points are
>> compatible.
>> At least, I never did it (i.e. used only unshifted k-vectors).
>>
>> As concerns your joke about August, well, it is pretty good. And add
>> September, October, November, etc.
>> This one recalled me a famous math problem ...
>>
>> Bests,
>> Eyvaz.
>>
>> -------------------------------------------------------------------
>> Prof. Eyvaz Isaev,
>> Theoretical Physics Department, Moscow State Institute of Steel & Alloys,
>> Russia,
>> Department of Physics, Chemistry, and Biology (IFM), Linkoping University,
>> Sweden
>> Condensed Matter Theory Group, Uppsala University, Sweden
>> Eyvaz.Isaev at fysik.uu.se, isaev at ifm.liu.se, eyvaz_isaev at yahoo.com
>>
>>
>> --- On Fri, 8/8/08, Miguel Martínez <miguel.martinez at ehu.es> wrote:
>>
>>
>>> From: Miguel Martínez <miguel.martinez at ehu.es>
>>> Subject: [Pw_forum] Silent Calculations
>>> To: "PWscf Forum" <pw_forum at pwscf.org>
>>> Date: Friday, August 8, 2008, 1:28 PM
>>> Dear everybody,
>>>
>>> I've been running into a few issues as of late, when
>>> running some
>>> calculations, it seems the calculations have stalled for a
>>> pretty long
>>> period of time.
>>>
>>> The one that troubles me less, taking into account that
>>> it's pretty
>>> expensive, is an electron-phonon calculation. This is a
>>> pretty large
>>> system (20 atoms in the primitive cell) and computing the
>>> phonons at
>>> gamma took ages (well two days using 4 processors). It
>>> would be great
>>> if some output were written just to remind you the code is
>>> still
>>> working, maybe something related to convergence. But I find
>>> somewhat
>>> disturbing that no line was written since the electron
>>> phonon
>>> calculation began.
>>>
>>> My other issue is a very similar stalling in a Raman
>>> calculation. No,
>>> it's not the same structure. Here, I lack from
>>> experience, since this
>>> is the first Raman calculation I've done outside the
>>> examples.
>>>
>>> Self-consistent input is:
>>> &control
>>> calculation='scf'
>>> restart_mode='from_scratch',
>>> prefix='sih4',
>>> tstress = .true.
>>> tprnfor = .true.
>>> pseudo_dir = '/home/wmbmacam/pseudo/',
>>> outdir='/scratch/wmbmacam/fdd2-raman/'
>>> /
>>> &system
>>> ibrav= 10, celldm(1) = 12.65898481, celldm(2) =
>>> 1.07100858,
>>> celldm(3) = 0.621208699, nat=10, ntyp= 2,
>>> ecutwfc = 55.0
>>> nr1= 72, nr2= 72, nr3= 72
>>> /
>>> &electrons
>>> conv_thr = 1.0d-8
>>> mixing_beta = 0.7
>>> /
>>> ATOMIC_SPECIES
>>> Si 28.0855 Si.pbe-rrkj.UPF
>>> H 1.0079 H.pbe-vbc.UPF
>>> ATOMIC_POSITIONS {crystal}
>>> Si 0.614368447799 0.385631552201 0.614368447799
>>> Si 0.864368447799 0.635631552201 0.864368447799
>>> H 0.442372877617 0.836174882136 0.364427192728
>>> H 0.364427192728 0.357025047520 0.442372877617
>>> H 0.892974952480 0.885572807272 0.413825117864
>>> H 0.413825117864 0.807627122383 0.892974952480
>>> H 0.378768583324 0.338682246661 0.003482898532
>>> H 0.003482898532 0.279066271483 0.378768583324
>>> H 0.970933728517 0.246517101468 0.911317753339
>>> H 0.911317753339 0.871231416676 0.970933728517
>>> K_POINTS {automatic}
>>> 12 12 12 1 1 1
>>>
>>>
>>> The phonon input is:
>>>
>>> phonons of Sih4 at Gamma
>>> &inputph
>>> tr2_ph=1.0d-16,
>>> prefix='sih4',
>>> epsil=.true.,
>>> trans=.true.,
>>> lraman=.true.,
>>> elop=.true.,
>>> amass(1)=28.0855,
>>> amass(2)=1.00794,
>>> outdir='/scratch/wmbmacam/fdd2-raman/',
>>> fildyn='sih4.dynG',
>>> fildrho='sih4.drho',
>>> /
>>> 0.0 0.0 0.0
>>>
>>>
>>> The calculations stalled right after the effective charges
>>> calculations.
>>> After that, it spent 24 hours without writing anything (it
>>> needed 10
>>> hours to compute the effective charges). No convergence, no
>>> threshold,
>>> no crashes, no nothing.
>>>
>>> Is this behaviour normal? Or are my compiled binaries just
>>> telling me I
>>> should stop bothering them in August?
>>>
>>> Finally, I just wanted to note that the amasses in example
>>> 15 are
>>> seemingly wrong. While the masses in the scf run are fine,
>>> the ones
>>> used in the phonon calculations correspond to GaAs instead
>>> of AlAs.
>>>
>>> Regards,
>>>
>>> Miguel
>>>
>>> --
>>> ----------------------------------------
>>> Miguel Martínez Canales
>>> Dto. Física de la Materia Condensada
>>> UPV/EHU
>>> Facultad de Ciencia y Tecnología
>>> Apdo. 644
>>> 48080 Bilbao (Spain)
>>> Fax: +34 94 601 3500
>>> Tlf: +34 94 601 5326
>>> ----------------------------------------
>>>
>>> "If you have an apple and I have an apple and
>>> we exchange these apples then you and I will
>>> still each have one apple. But if you have an
>>> idea and I have an idea and we exchange these
>>> ideas, then each of us will have two ideas."
>>>
>>> George Bernard Shaw
>>>
>>> _______________________________________________
>>> Pw_forum mailing list
>>> Pw_forum at pwscf.org
>>> http://www.democritos.it/mailman/listinfo/pw_forum
>>>
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://www.democritos.it/mailman/listinfo/pw_forum
>>
>>
>>
>
>
> ************************************************************************
> Dr. Subhradip Ghosh
> Assistant Professor
> Department of Physics
> Indian Institute of Technology
> Guwahati,Assam-781039
> India
> E-mail:subhra at iitg.ernet.in
> Phone: +91 361 2582717(O)
> +91 361 2584717(R)
> Fax: +91 361 2582749 (Physics)
> +91 361 2690762 (General)
> ************************************************************************
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
More information about the users
mailing list