[QE-developers] Different magnetisation values from version 7.3.1 when compared with 6.7

[Pietro Davide Delugas]

P.S.
The convergence threshold 1.d-6 is high, thus a bit prone to having random final densities and magnetisations.
The starting magnetisations in 6.7, especially for Ni, are colossal: 1.0, used to mean that all valence electrons are in the majority channel. The situation has changed in newer versions, even if I don't remember if it was already like that in 7.3.1, probably yes. Values less than 1 are considered as before as fractional magnetisation (so total magnetisation of the ion divided by the total valence charge of the ion). For values equal to or larger than 1.0, this number now represents the total magnetisation of the ion, so the same inputs but different starting magnetisation in the two versions.
pietro

________________________________
From: developers <developers-bounces at lists.quantum-espresso.org> on behalf of Jaime Oliveira da Silva <jaime.silva at fis.uc.pt>
Sent: Tuesday, June 24, 2025 14:23
To: developers at lists.quantum-espresso.org <developers at lists.quantum-espresso.org>
Subject: [QE-developers] Different magnetisation values from version 7.3.1 when compared with 6.7

Hi everyone,

I want to report a regression bug when doing the calculation of Nickel slab. In qe version 6.7 the values for the slab magnetisation is different than the version 7.3.1. However, the convergence is faster in the more recent version when compared with version 6.7. Is this a bug or there is something new in version 7.3.1?

I attached the input file of the calculation.

Cheers,
Jaime Silva
Junior Researcher

Universidade de Coimbra • Faculdade de Ciências e Tecnologia • Departamento de Física • CFisUC
University of Coimbra • Faculty of Sciences and Technology • Department of Physics • CFisUC








-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20250625/f70bb262/attachment-0001.html>