Yambo Community Forum

I see, thank you!

Dear all,

Do we need to run momentum-dependent RPA or BSE calculations in the full Brillouin zone instead of IBZ?
And are the q-vectors calculated only for the first Brillouin zone?

Best

Dear all,

You can find dielectric functions for q-dependent ipa and bse calculations with or without taking into account the symmetries in the system.
After calculating the ground state I switched-off the symmetries with ypp.
It seems that ipa for a specific momentum is giving the same results with or without symmetry while bse is giving very different spectra.
Does anyone knows what is wrong?

Best

Dear Fadıl,

actually, we should investigate what's going wrong here and from the output file only it is not possible to understand.

Can you post the report and input files for the two calculations? I do not understand how you have the same dimension of the excitonic matrix independently of the usage of the symmetries.

In case we would need to reproduce your problem, it would be preferable if you could update to the newest release of the code and repeat the calculation.


Best,
Daniele

Hi Daniele,

Thank you for the response. Unfortunately, I am having some issues to install the latest version of the code to our cluster.
But I did the similar calculations in an another version of the code (5.0.3) and they gave the same issue.
You can find input, output and report files (with symmetry and without symmetry (labeled as no_sym) ) in the attachment.

Best regards

Dear Fadil,
Thanks for reporting, we will investigate the issue.
It is possible we will have to reproduce your problem, in this case I will ask you QE input and pseudo. It would be desirable to have a minimal (low parameter, fast calculation) example showing the problem.

Best,
Daniele

Dear Fadil,
this is a known issue for simulations with SOC.
Indeed, in the gpl, finite-q bse with SOC has been switched off.

There is a branch which you can use where this issue has been fixed.
However the bse implementation there is numerically a bit more demanding.
https://github.com/yambo-code/yambo/tre ... nite-q-dir

The issue exists only in presence of SOC.

Best,
D

Dear Daniele,

Since this is a known issue I assume there is no need to reproduce the problem otherwise please let me know.

Dear Davide,

Thank you for the clarification.
Then, SOC-included finite-q bse calculations by considering IBZ or full-BZ (switching-off symmetries by ypp) have problems in both?

Best

Yeah, both simulations with and without symmetries have issues. My suggestion is to check the branch with the fix.

Best,
D

Thank you!