Hello all,

While this calculation was successfully converged in 5.0_develop_branch, it seems that it never converges in 5.1_release with the same setting. I'd like to attach two inp.xml files from both versions. And also the convergence log is attached for 5.1 one. The number of iterations in the file is 50 but it was run several times. I'd appreciate if anyone could provide some advice for successful convergence. I know that there were some updates with regards to non-colliearity between 5.0 and 5.1. But at this moment I don't know how to proceed further.

Cheers,
Wonjae

Dear Wonjae,

indeed I believe there is a bug in the unconstrained non-collinear magnetism in the 5.1 release. It is hopefully fixed in the recent development branch.

Daniel

Dear Daniel

Thank you for your answer, I just read the documentation on non-collinear magnestism. It says at the end that it is helpful to adjust

calculationSetup/nocoParams/@mix_RelaxWeightOffD

to be the same as the "usual mixing parameter" used for the system. Does this parameter mean the one present in

<ldaU l_linMix="F" mixParam=".050000" spinf="1.000000"/> ??

The description is in the last paragraph of "Fleur homepage - User guide - Theoretical background - Treatment of noncollinear magnetism" (Cannot post an external link)

Cheers
Wonjae