[QE-users] Result precision across different builds

[Nicola Marzari]

Dear all,

there are cases (typically, magnetic systems, especially with hybrids or dft+u) where there are different local/global minima, and one can randomly converge in one or the other. All are numerically correct self consistent states.

And there are rare cases where, if no noise is there at start (that’s why good to use atomic plus random) some eigenstates at some kpoints are missed, leading to an incorrect minimum.

Otherwise, there shouldn’t be meaningful differences, especially if one dies a restart with one/few iteration(s) after a very large cell relaxation, to account for the change of volume and the consequent change of plane wave basis.

HTH

nicola

Sent from a tiny keyboard... Contact info:
http://theos-wiki.epfl.ch/Main/Contact

On 28 Aug 2026, at 09:04, 박기명 <knppkm at naver.com> wrote:


In very rare cases, setting &ELECTRONS startingwfc = 'random' can lead to a different converged value. Have you checked that the inputs are identical?
Also, did you set conv_thr tight enough? The default value is not good enough. And in my case, I only trust the total energy down to the sixth decimal place. The unit is eV.

Best regards,
GiMyung Park
Yonsei University

-----Original Message-----
From: "Simon Michnowicz via users"<users at lists.quantum-espresso.org>
To: <users at lists.quantum-espresso.org>;
Cc:
Sent: 2026-08-28 (금) 10:52:09 (GMT+09:00)
Subject: [QE-users] Result precision across different builds


Dear User List
I am helping a researcher install Quantum Espresso. I find different combinations of compilers/libraries produce slightly different results upon a sample input given to me. The researcher said the results should be consistent up to 5 decimal points.  Is the  variation that I see to be expected, or could there some other problem here?

Tool                                       Final Energy
Intel Compiler,MKL               -314.1147158171
gcc/15.2.0-gcc-11.5.0/MKL  -314.1256316418
gcc/11.4/openblas                -314.1256341822

much thanks

---
Simon Michnowicz
Senior Application Specialist,  High-Performance Computing

Monash eResearch Centre
Monash University
15 Innovation Walk, Building 75, Clayton Campus
Wellington Road, VIC 3800
Australia

T:  +61 3 9902 0794
M: +61 3 0418 302 046
E: simon.michnowicz at monash.edu<mailto:simon.michnowicz at monash.edu>
monash.edu<http://monash.edu/>
[https://mail.naver.com/readReceipt/notify/?img=3seR1XImW4Ropx%2BCFAmsFrF0F4M%2FKqbdazUlpxEqFAgXKrMZpxvZpxiCtzFXp6UmKxg5W4dQaXkqpBiqtzwGbX3q7NJ9M6ld%2B6dRpBFmb4kqbrY5WXiN.gif]
_______________________________________________________________________________
The Quantum ESPRESSO Foundation stands in solidarity with all civilians worldwide who are victims of terrorism, military aggression, and indiscriminate warfare.
--------------------------------------------------------------------------------
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users at lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20260828/a19b2957/attachment.html>