oscillatory hydock

Please post here any issue related to convergence parameters in linear response GW and BSE calculations.

Moderators: Davide Sangalli, Daniele Varsano

Post Reply
burkzdemir
Posts: 47
Joined: Tue Sep 09, 2014 7:57 pm

oscillatory hydock

Post by burkzdemir » Mon Feb 25, 2019 5:41 pm

Dear Yambo developers,

I am running GW+BSE calculation. I was able to run with "-y d" option and results are smooth however when I run with "-y h" results are oscillatory. What can I do to make the results smooth (flat) with "-y h" option which is much more faster as I experienced.

Burak,
Best
Burak Ozdemir
Post-doc,
University of Ostrava, Ostrava, Czech Republic

User avatar
Daniele Varsano
Posts: 2374
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: oscillatory hydock

Post by Daniele Varsano » Mon Feb 25, 2019 6:00 pm

Dear Burak,
you can surely make your threshold more stringent: a lower value for

Code: Select all

BSHayTrs
.
If it does not solve your problem please post an example of what you are observing.

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/

haseebphysics1
Posts: 129
Joined: Sat Aug 17, 2019 2:48 pm

Re: oscillatory hydock

Post by haseebphysics1 » Sat Sep 07, 2019 4:04 pm

Dear Daniele,

I had reduced MaxGVecs from around 104000 to 52000 RL vectors using -V RL. (I'm doing just BSE)
Now after converging static screening parameters I'm going to converge macroscopic dielectric parameters for BS kernel which also includes BSENGexx. In my case, the value for BSENGexx is close to my reduced G vectors?

Code: Select all

BSENGexx=  51867       RL
Is it okay to go with this number and consider it as converged? Or do I need to vary it as well to see the convergence?
The value of this parameter in tutorials was varied as 10, 20, 30 Ry etc!

And secondly what is the difference between RL and Ry units and what is the conversion factor?

Regards,
Haseeb Ahmad,
MS - Physics,
LUMS, Pakistan
Haseeb Ahmad
MS - Physics,
LUMS - Pakistan

User avatar
Daniele Varsano
Posts: 2374
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: oscillatory hydock

Post by Daniele Varsano » Sun Sep 08, 2019 10:05 am

Dear Haseeb,
please post your question in the right threads. If you do not find an already thread addressing your question you can open a new one.
Your post now is in a thread discussing the oscillatory haydock and doe not fit. I tell you this because an organized forum facilitate the users to find useful info on their problem.
BSENGexx needs to be converged, anyway this part it is not particularly time-consuming so you can consider starting with a large value directly. What you can do is to reduce the number and verify you do not have a great change in the final result.
And secondly what is the difference between RL and Ry units and what is the conversion factor?
RL: Number of reciprocal lattice vectors
Ry: kinetic cutoff in energy (Rydberg), this will include a number of G vectors such that G^2/2 < Ecut

So this number will depend on the size of your supercell.

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/

Post Reply