How to iterate the wave function when calculating GW

You can find here problems arising when using old releases of Yambo (< 5.0). Issues as parallelization strategy, performance issues and other technical aspects.

Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano, Conor Hogan

Locked
Quxiao
Posts: 96
Joined: Fri Mar 26, 2021 11:27 am

How to iterate the wave function when calculating GW

Post by Quxiao » Thu Aug 25, 2022 5:01 pm

Dear Daniele:
Recently,I have new a new question after I read some paper about the GW calculation.Generally speaking, when we calculate the GW, we use KS state to approximate and replace the wave function, which does not cause great difference. However, when we study special problems, I saw that yambo introduced the iteration of eigenvalue when calculating GW, but how to make the eigenvalue and wave function iterate when calculating GW. The main reason is that I haven't found a method to calculate the iterative wave function of GW, which is what I am confused about at present. Could you give me some help?
Best wishes
Quxiao
Quxiao in BIT,calculate the exciton

User avatar
Daniele Varsano
Posts: 3773
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: How to iterate the wave function when calculating GW

Post by Daniele Varsano » Mon Aug 29, 2022 8:43 am

Dear Quixiao,

when you calculate the self-energy self-consistently, the easiest way is the eigenvalue only self-consistency. This can be done by iterating both G and W or G only as explained here:
https://www.yambo-code.eu/wiki/index.ph ... alues_only

In this case, as you can easily understand the wfs are not changed, ie KS wfs are used during all the procedures.

If you want to perform a self-consistent calculation while also updating the wfs, you can do it at COHSEX level (not GW). This is done by using the yambo_sc executable. You can have a try, this feature anyway has been released recently and has not been extensively documented.

Best,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/

Locked