Yambo 5.0.4 parallization strategy for large system [nbnds =3600]

[harrier_class]

Dear Dr. Varsano,

I started my TDDFT calculations in the transition space while neglecting non-local commutator effect. It goes nicely, and I have a spectrum. Thanks for all your suggestions.

Since, I wanna compare the spectra with the spectra obtained while keeping the non-local commutator effect, I would be interested in knowing if there is any way of making the dipoles calculations faster?

Just to mention, I am using Yambo 5.0.4, and we are interested in the spectra in the energy range of 0 to 6 eV.

Further, I came across the option called 'DipBands'. As my system has around 2750 filled bands, and my nscf calculations are done using 4000 total bands, and we are interested in low energy spectra, I am under the impression that changing the Dipbands, will make the calculations much faster.

Code: Select all

% DipBands
1 | 30 | # [DIP] Bands range for dipoles
%

Would you recommend me to change this, and do the convergence study for the same?

Thanks a lot for all your support.

--
Best regards,
Vipul

[Daniele Varsano]

Dear Vipul,
the inclusion of the non local commutator is computationally intensive.
You can set the DipBands range equal to the bands range you use in the calculation BSEBands.

Best,

Daniele

[harrier_class]

Dear Dr. Varsano,

I appreciate all your support along the way.

The calculations went through smoothly.

Have a nice rest of the day.

--
Best regards,
Vipul