Page 1 of 1

Bug reports for low scaling full self-consistent GW calculations

Posted: Wed May 15, 2024 1:25 am
by zwxiao7
Dear Merzuk Kaltak,

Something wrong when I send you emails. In case of you haven't receive my email, I post it here.

The problem I meet (Just the same as I have said in my letters):
"As you have said in the VASP wiki web site “the output of the GWR is similar to partially self-consistent GW calculations, with the difference that KS-energies are replaced by the QP energies from previous iteration”, but when I use this method to calculate the band gap or band edges of CsSnCl3 (Pm-3m), I found the KS-energies are not replaced by the QP energies during the all iterations. I got the information of the KS-energies and QP energies from the OUTCAR file. So, there maybe two problems evolved, firstly, the OUTPUT messages are wrong, it doesn’t OUTPUT the correct messages; secondly, the iterations are wrong, it doesn't use the renormalized QP energy to iterate the G. Although I have repeated this action several times, in case of it is an operational problem, I have attached my INCAR and OUTCAR files."

Thank you for your suggestions about the vertex corrections in GW calculations.

Best,
Yongcheng.

Re: Bug reports for low scaling full self-consistent GW calculations

Posted: Wed May 15, 2024 8:10 am
by merzuk.kaltak
Dear Yongcheng,

I think your problem is resolved by switching from ALGO=GWR to ALGO=EVGW0R.
The former solves the Dyson equation for G and W, while EVGW0R does only a pole-shift of G.