[QE-developers] ESM examples does not reproduce the reference results

[大谷実]

Dear Javad and Paolo,

Thank you for testing the ESM examples. I have reproduced Paolo’s result on my MacOS (gfortran, FFTW3, Open-mpi). Regarding, the discrepancy between "results" and "reference" in large |z| (>5), I put wrong esm1 files in the “reference” directory. So current results are ok for bc3. This difference is just a cosmetic one which does not affect any results.

The results for "pbc" cases are entirely wrong. I will check the origin of the error.

Javad:
I plotted your esm1 file; “Al001_bc1-reference.esm1” and “Al001_bc1-with6.3-rc2.esm1”. The third column (v_hartree) is different from the reference. As you mention, the results might depend on FFT. The problem originated from a FFT library is very complicated, so it is difficult for me to solve this.

As long as you can obtain the same results with the reference, probably your results are reliable. But please check the esm1 file carefully whether physically relevant potential profiles are obtained.

Best,
Minoru

--------------------------------------------------------------------------------------------------------
Minoru Otani
National Institute of Advanced Industrial Science and Technology,
Research Centre for Computational Design of Advanced Functional Materials
E-mail: minoru.otani at aist.go.jp<mailto:minoru.otani at aist.go.jp>
--------------------------------------------------------------------------------------------------------



From: developers <developers-bounces at lists.quantum-espresso.org> on behalf of Seyed Javad Azimi <j.azimi at email.kntu.ac.ir>
Reply-To: General discussion list for Quantum ESPRESSO developers <developers at lists.quantum-espresso.org>
Date: Thursday, June 28, 2018 17:18
To: developers <developers at lists.quantum-espresso.org>
Subject: Re: [QE-developers] ESM examples does not reproduce the reference results

Thanks Paolo for quick reply and attention.

Yes, the discrepancy reduces to pbc case (also a bit for bc3, as mentioned by Paolo) when FFT library is included.
(My first execution was on a clean virtual Ubuntu 16.04 only with gfortran compiler + qe internal LAPACK/BLAS without FFT & MPI )
It seems that it depends on FFT, however for the pbc case inconsistency is still there.

Does it challenge the reliability of the results ?!

Thanks for all comments.
Javad

________________________________
From: "Paolo Giannozzi" <p.giannozzi at gmail.com>
To: "developers" <developers at lists.quantum-espresso.org>
Cc: "Seyed Javad Azimi" <j.azimi at email.kntu.ac.ir>
Sent: Thursday, June 28, 2018 9:27:24 AM
Subject: Re: [QE-developers] ESM examples does not reproduce the reference results

Same results in serial execution and with a different compiler. There are also some differences for the case bc3 at large |z| (> 5).
Paolo

On Wed, Jun 27, 2018 at 2:40 PM, Paolo Giannozzi <p.giannozzi at gmail.com<mailto:p.giannozzi at gmail.com>> wrote:
This is what I get, and it seems ok. I found differences with reference .esm files only for the pbc case. Don't know why.
Paolo

On Wed, Jun 27, 2018 at 12:33 PM, Seyed Javad Azimi <j.azimi at email.kntu.ac.ir<mailto:j.azimi at email.kntu.ac.ir>> wrote:
Dear Developers,

I have noticed that the ESM example in qe/PW and qe/NEB does not reproduce the reference results for calculated potentials in .esm files and the divergency is huge.
The reference results were calculated with 6.1 but the next releases of qe (also 6.3-rc2) does not produce compatible results with reference .esm files (the .out files are compatible).
I have checked with various compilers and libraries and here I have attached one executed with gfortran + internal BLAS/LAPACK.

Maybe it needs a small fix !

Best,
Javad

--
Seyed Javad Azimi
PhD student | Physics Department
K. N. Toosi University of Technology, Tehran, Iran.
Università di Modena e Reggio Emilia, Modena, Italia.

--
This message has been scanned for viruses and dangerous content by
KNTU Antispam System (E.F.A. Project)<http://www.efa-project.org>, and is believed to be clean.

_______________________________________________
developers mailing list
developers at lists.quantum-espresso.org<mailto: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


--
Seyed Javad Azimi
PhD student | Physics Department
K. N. Toosi University of Technology, Tehran, Iran.
Università di Modena e Reggio Emilia, Modena, Italia.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20180628/97f6c93b/attachment-0001.html>