[CP2K-user] [CP2K:21932] Inquiry about SCF Convergence Issues with the Diagonalization Method in CP2K

[Johann Pototschnig]

You can try to increase the cutoffs:
    &MGRID
      CUTOFF  400
      REL_CUTOFF  55
    &END MGRID

I t can be  helpful to decrease
 ALPHA
or change the buffer
NBROYDEN

On Tuesday, October 28, 2025 at 7:27:57 AM UTC+1 Que Zhu wrote:

> I reviewed your output file and found that Convergence has been 
> oscillating.
>
> yis... at 163.com <yis... at 163.com> 于2025年10月24日周五 21:12写道：
>
>> Dear CP2K Developers and Experienced Users,
>>
>> Hello everyone,
>>
>> I am a beginner with CP2K and am encountering an issue with 
>> Self-Consistent Field (SCF) convergence when using the diagonalization 
>> method, even after confirming the correct spin multiplicity for my system.
>>
>> To identify the ground state, I first performed a spin multiplicity scan 
>> using the Orbital Transformation (OT) method. The results confirmed that a 
>> spin multiplicity of 17 yields the lowest energy for my system.
>>
>> Subsequently, I switched to the diagonalization method with k-points to 
>> perform a geometry optimization. However, I found the SCF procedure to be 
>> very difficult to converge under this setup.
>>
>> I am writing to seek your advice and have attached my input and output 
>> files for reference. Specifically, I would like to ask:
>>
>> Are there any issues with my input file, and is there room for further 
>> optimization?
>>
>> Why does the diagonalization method often struggle with convergence 
>> compared to the OT method?
>>
>> Under what conditions can the convergence speed of the diagonalization 
>> method be comparable to that of the OT method?
>>
>> Any insights, suggestions, or guidance you could provide would be greatly 
>> appreciated. Thank you for your time and help.
>>
>> Best regards,
>>
>> Yisi chi 
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "cp2k" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to cp2k+uns... at googlegroups.com.
>> To view this discussion visit 
>> https://groups.google.com/d/msgid/cp2k/0d79829a-8a88-4363-adee-a7cb741d9e16n%40googlegroups.com 
>> <https://groups.google.com/d/msgid/cp2k/0d79829a-8a88-4363-adee-a7cb741d9e16n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/fab6d7bb-5ff6-44a3-95b7-7138f311fa3cn%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20251028/85afd3b0/attachment.htm>