GW precision

Concerns issues with computing quasiparticle corrections to the DFT eigenvalues - i.e., the self-energy within the GW approximation (-g n), or considering the Hartree-Fock exchange only (-x)

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

Post Reply
sohan
Posts: 32
Joined: Thu Nov 14, 2024 1:10 pm

GW precision

Post by sohan » Fri May 23, 2025 12:21 pm

Dear Sir,
Do we need to do G0W0 calculation in single precision or double precision?
regards
sohan
Sohan
M.Sc. (Physics)
Central University of Punjab, India

User avatar
Daniele Varsano
Posts: 4223
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW precision

Post by Daniele Varsano » Sat May 24, 2025 7:26 am

Dear Sohan,

considering the overall precision of the method, single precision is the default and it is usually enough.

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/

sohan
Posts: 32
Joined: Thu Nov 14, 2024 1:10 pm

Re: GW precision

Post by sohan » Wed May 28, 2025 7:28 pm

Dear Sir,
I have done some g0w0 in double precision, can I continue with double precision (I know it'll take more time than single precision)?
regards
sohan
Sohan
M.Sc. (Physics)
Central University of Punjab, India

User avatar
Daniele Varsano
Posts: 4223
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW precision

Post by Daniele Varsano » Thu May 29, 2025 7:26 am

Dear Sohan,

yes, sure you can do calculations in double precision. In this way you will have a better accuracy but calculation are more memory consuming.

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/

sohan
Posts: 32
Joined: Thu Nov 14, 2024 1:10 pm

Re: GW precision

Post by sohan » Thu May 29, 2025 10:45 am

Dear Sir,
Thank you for the response. I have used the yambo_nl executable in g0w0 for some material instead of yambo. :| Will this create any discrepancy in output? Eagerly waiting for your kind response.
regards
Sohan

Note: I have tried with both the executable on the hBN example available on the wiki, and they give the exact same results of g0w0. So I shouldn't be worried about using yambo_nl for g0w0?
Sohan
M.Sc. (Physics)
Central University of Punjab, India

User avatar
Daniele Varsano
Posts: 4223
Joined: Tue Mar 17, 2009 2:23 pm
Contact:

Re: GW precision

Post by Daniele Varsano » Fri May 30, 2025 7:43 am

Dear Sohan,

Sure, you can also use the yambo_nl for g0w0 as they are compatible. Even if it should be safe, anyway my suggestion is to use the yambo executable for further calculations.

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