Yambo Community Forum

Dear all,
Can I use double-grid method in the calculations of nonlinear properties(such as TD-BSE and second harmonic generation)?
Thanks in advance.

No, for the moment it is not possible

best
Claudio

Dear Claudio,
Thanks for your reply.
So, the k grid should be tested carefully when calculating the SHG of 2D materials?
In the papers (PRB,3,074003,2019), authors set a dense k-point, 24x24x1 for InSe but only 200 bands for GW-BSE calculations.
I want to know that k-point sampling is more important than the number of bands in GW-BSE or SHG calculations?

Dear Yimin Ding
a large number of bands is necessary to converge the dielectric constant
that enter in the definition of W and for the calculation of the GW corrections.
For the non-linear response, and the BSE you just need a small number of conduction and valence bands,
while convergence in k-point is more difficult.

My advice is to converge first the independent particle case,
and then use the same number of k-points in the GW and BSE.
For the number of conductions bands to use, it is best to test it in the linear response case that is faster

best
Claudio

Dear Claudio,
Thanks for your reply.
I am not clear about this sentence, “My advice is to converge first the independent particle case, and then use the same number of k-points in the GW and BSE.”
Do you mean convergence test of k-point in pw.x code first? I think that would be a very fast test.

Dear Yimin Ding,
with the new release (yambo 5.0) it is now possible to use the double-grid approach in real-time simulations.

Best,
D.

Dear Davide,
Many thanks.
But, I wonder how to perfom double-grid approach in real-time simulations? Is it the same as that in common BSE calculations?