[QE-developers] Incompatible FFT grid in ph.x calculation

[Xiao Zhang]

Dear Paolo,

I see. Thanks for checking this and I will try what you suggested. So in
general is there a good number of digits that we should shoot for? Or is it
more system dependent, and we just need to play around it?

Thanks!
Best!
Xiao

On Fri, Aug 14, 2020 at 9:37 AM Paolo Giannozzi <p.giannozzi at gmail.com>
wrote:

> I see: it's the usual problem with "almost-symmetric" coordinates. If you
> transform the "almost 1/3" and "almost 2/3" into something closer to 1/3
> and 2/3 in the x and y atomic coordinates, the error disappears. In the
> previous version of the code, the checks were done  in a slightly different
> way and this problem did not show up.
>
> Paolo
>
> On Fri, Aug 14, 2020 at 3:22 PM Paolo Giannozzi <p.giannozzi at gmail.com>
> wrote:
>
>> On Fri, Aug 14, 2020 at 2:51 PM Xiao Zhang <xiaozha at umich.edu> wrote:
>>
>> I realized that in 6.6, an additional line shows: Fft bands division:
>>> nmany = 1 (line 15). Not sure if this means anything or not?
>>>
>>
>> it doesn't.
>>
>> Apparently the check on fractional translations being commensurate with
>> the real-space grid is a little bit too strict. Not sure how to fix it in a
>> safe way, though
>>
>> Paolo
>>
>>
>>> Thanks!
>>> Best!
>>> Xiao
>>>
>>> On Fri, Aug 14, 2020 at 8:31 AM Paolo Giannozzi <p.giannozzi at gmail.com>
>>> wrote:
>>>
>>>> It would be useful to see the scf output for v.6.5 as well
>>>>
>>>> Paolo
>>>>
>>>> On Fri, Aug 14, 2020 at 9:14 AM Xiao Zhang <xiaozha at umich.edu> wrote:
>>>>
>>>>> Dear QE developers,
>>>>>
>>>>> I'm sending the email asking about an error I get from the phonon
>>>>> calculations in the 6.6 version of quantum espresso. I'm trying to run
>>>>> phonon calculation for 2H SiC and have the scf run finished with no
>>>>> problem, but when I run phonon calculation following that, I received an
>>>>> incompatible FFT grid error. This happened only with the new version of QE
>>>>> while using the exact same input, QE 6.5 works fine and provides output. I
>>>>> was wondering if this is something we expect? (I do realize there is one
>>>>> potentially relevant tag on the release note under "fixed")
>>>>>
>>>>> I attached one example just in case it might help: This is 2H-SiC and
>>>>> I attached the input and output from scf, input and output from ph.x from
>>>>> both 6.5 and 6.6. The k-points and cutoffs are reduced for quick tests, but
>>>>> converged k-points and cutoffs show the same issue. The structure is
>>>>> relaxed within QE, and the PP used is ONCV-PBE. I'm a bit confused why this
>>>>> is seen in phonon calculation but not the scf calculation? I assumed that
>>>>> if it's a problem of my input structure, I should've seen a problem in
>>>>> both? Any hint or suggestions would be very helpful!
>>>>>
>>>>> Thanks a lot!
>>>>> Best!
>>>>> Xiao Zhang
>>>>> Department of Materials Science and Engineering
>>>>> University of Michigan, Ann Arbor
>>>>>
>>>>> _______________________________________________
>>>>> developers mailing list
>>>>> developers at lists.quantum-espresso.org
>>>>> https://lists.quantum-espresso.org/mailman/listinfo/developers
>>>>>
>>>>
>>>>
>>>> --
>>>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>>>> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
>>>> Phone +39-0432-558216, fax +39-0432-558222
>>>>
>>>>
>>
>> --
>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
>> Phone +39-0432-558216, fax +39-0432-558222
>>
>>
>
> --
> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20200814/7890570d/attachment-0001.html>