How to determine the size of the FFTGvecs?

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
gaohan
Posts: 15
Joined: Sun Feb 07, 2021 3:56 am

How to determine the size of the FFTGvecs?

Post by gaohan » Thu Mar 11, 2021 3:37 am

Dear Developers,
I know the FFTGvecs is the number of G-vectors or energy cutoff for expanding wavefunctions in transition matrix elements and FFT operations, and adjusting that can regulate the memory of calculation. But I wonder how to determine it?
For example, I do the DFT calculation using 450eV for the first step. Is that meant to be that I should set the FFTGvecs to 450 eV?
The default value is 15529 RL when I generate the yambo input file.
Looking forward to your reply.
Han Gao
School of Physics, State Key Laboratory of Crystal Materials, Shandong University, Jinan 250100, Shandong, China

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

Re: How to determine the size of the FFTGvecs?

Post by Daniele Varsano » Thu Mar 11, 2021 9:52 am

Dear Han Gao,

if you are not experiencing memory problem you can leave the default value.
If you need to lower it, a reasonable criteria is to look at the orthonormalisation of the wfs which is reported when calculating e.g. dipoles or HF self-energy:
e.g.

Code: Select all

[WF-Oscillators/G space loader] Normalization (few states)  min/max  :0.260E-09  1.00
and verify that it does not differ from 0,1 which means <n|n'>=delta_nn'

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