Strange in my sum-over-states calculation

[lyzhao]

Dear all,
I try to code the following formular (from Yambo-Cheatsheet-5.0) to calculate the EPS, as a test, I use the two-dimensional BN and do not use coulomb cutoff

eps.png

First I run an ipa calculation to obtain the optical matrix element <v|r|c>,
then calculate <0|r|lamda> to obtain BSE dipoles, finally calculate the EPS.
I got the strange results when comparing my results with Yambos as follow

vs.png

that is, for k3 mesh of 3x3x1, my results are in good agreement with yambo,
however for k6 mesh of 6x6x1, clear differences are observed and it seems that the main difference is the intensity.

Is there k-dependent parameter I have missed?

Any help will be appreciated.

[Daniele Varsano]

Dear Youzhao,

hard to say what is missing.
I suggest you to check the number of K points in the BZ, that are obtained by applying the symmetries to the k points in the IBZ.
According to the report files you should have 3 kpoints in IBZ and 9 in the BZ for the 3x3 and 7 kpoints in IBZ and 36 in BZ for the 6x6.

Best,
Daniele