Page 1 of 1

DysSolver = "s" does not work

Posted: Fri Apr 01, 2022 5:37 pm
by anhhv
Dear YAMBO communitiy,

I'm running GW with YAMBO 4.4.0. I tried to set DysSolver= "s" but in the output it was changed to "n". I don't know why. Looking at the source code, I found that this only could be changed in src/interface/INIT_barriers.F

if ( (l_ppa.and.trim(QP_solver)=='s').or.l_cohsex ) QP_solver='n'


However, I checked l_ppa and l_cohsex in src/modules/mod_drivers.F and they're all set FALSE.

Please give me some clues. I attached here my input file.

Re: DysSolver = "s" does not work

Posted: Fri Apr 01, 2022 11:49 pm
by Daniele Varsano
Dear Viet-Anh Ha,

I had a look at your input file, you are doing a gw calculation in plasmon pole approximation, as also stated from the name you gave to the file (ppa).
When using the plasmon-pole appoximation the secant method cannot be appkied. That's the reason of the barrier.

Best,
Daniele

Re: DysSolver = "s" does not work

Posted: Sat Apr 02, 2022 4:41 am
by anhhv
Hi Daniele,

Thank you for explaining. So, that means l_ppa was assigned to TRUE somewhere on-fly.

Viet-Anh

Re: DysSolver = "s" does not work

Posted: Sun Apr 03, 2022 5:51 pm
by anhhv
Hi Daniele,

I'm curious why with plasmon approximation being used, DysSolver = "s" was not interested. Is there any technical issue here?

Best,
Anh

Re: DysSolver = "s" does not work

Posted: Mon Apr 04, 2022 9:29 am
by Daniele Varsano
Dear Viet-Anh Ha,

it's because you need a reliable frequency dependency of the self-energy and PPA it is not.
Best,
Daniele