NaN in TDDFT calculation with yambo 4.2.1

[javad]

Dear people of yambo

I get NaN when I perform ALDA with yambo on a simple molecule
here my input files

with the best regards
Javad

[Daniele Varsano]

Dear Javad,
can you also add the report file of your calculation?
In any case my suggestion is to evaluate the TDDFT ALDA in configuration space more than in G space, as in G space you can have instability in the region of low density.
In order to do that you need to build the input as:
yambo -o b -k alda -y d
In this way you have also the advantage to include automatically all the G vectors in the kernel.
Best,
Daniele

[Jiuyusun]

Hi Daniele,

I am interested in:

In any case my suggestion is to evaluate the TDDFT ALDA in configuration space more than in G space, as in G space you can have instability in the region of low density.

I notice you had said this kind of suggestion several times in the form. Why ALDA in G space is less stable? Can you refer me some references about this issue?

Thanks
Jiuyu

[Daniele Varsano]

Dear Jiuyu Sun,
for a long time now Yambo has been using the libxc libraries to evaluate the kernel, and therefore I would say that instabilities problem are no longer present. Main difference in the two approaches is that in G-space you include a number of kernel components equal to the block considered in the response function, while in transition space the kernel is evaluated in real space and all the components are automatically considered.

Best,
Daniele