[QE-developers] constrained atom takes geometry steps with DAMP

[Katie Schwarz]

Please delete this report.  QE's temporary files were selectively
overwritten during the run, causing the problems described above.

On Thu, Jun 20, 2019 at 11:18 AM Katie Schwarz <kaschwar at gmail.com> wrote:

> Dear QE Developers,
>
> I am running a constrained optimization of a slab, with two constrained
> atoms, using DAMP, but the constraint does not consistently hold.
>
> In my example (attached), the constraint on one of the two atoms holds
> during many geometry steps.  The constraint on the other atom holds for 9
> iterations, and then moves on the 10th.  I have run similar jobs on
> multiple architectures now, and the pattern remains- one atom stays fixed
> by the constraint, and the other does not always.
>
> Here is a grep of the position of the copper atom that moves:
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.625170579    0   0   0
>
> NCu      0.500000000   0.500000000   0.624738221    0   0   0
>
>
> (I originally thought that there was a symmetry constraint that was
> causing the problem, which is why I created atoms called "NCu", the Cu
> atoms I'm constraining.  )
>
>
>  I have attached the input file + pseudopotentials, and the output file.
>
> As you can see from the output file, this is for PWSCF v.6.3, and a
> single node job with 64 processors, but this error also occurs on other
> clusters, and parallelized across multiple nodes.
>
> Thanks, and please let me know if additional information would be helpful.
>
> Best,
> Katie
>
> Katie Schwarz
> Research Chemist
> National Institute of Standards and Technology
>


-- 
Katie Schwarz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20190621/556767f2/attachment.html>