[ERROR] The whole BSE has zero dimension

[Zafer Kandemir]

Dear Developers,

Code: Select all

 <05m-07s> P1-s204: [WARNING] Null BSE kernel block dimension(s) found.
P1-s204: [ERROR] STOP signal received while in[06] Bethe Salpeter Equation @q1
P1-s204: [ERROR] The whole BSE has zero dimension !

I got an error like this and I don't understand why, where am I making a mistake?
I attached input, output and report files.

Best regards,
Zafer

[Daniele Varsano]

Dear Zafer,

your electronic structure is a metal with:

Code: Select all

  [X] Filled Bands                                  :  24
  [X] Metallic Bands                                :  25  28
  [X] Empty Bands                                   :    29   450

and in your BSE input file you set:

Code: Select all

% BSEBands
    45 |  56 |                       # [BSK] Bands range
%

which are all empty states, so no possible transitions are found.

Is it possible you switched from a spinor to a non-spin polarized calculation maintaining the same input file? In this case, the bands should be halved.

Best,
Daniele

[Zafer Kandemir]

Dear Daniele,

Thank you very much for your quick answer.
Since this is the case in the r_setup file, I set the bands in that range.

Code: Select all

r_setup:
  [X] === Gaps and Widths ===
  [X] Conduction Band Min                           :  11.57463 [eV]
  [X] Valence Band Max                              :  11.57463 [eV]
  [X] Filled Bands                                  :   48
  [X] Metallic Bands                                :   49   54
  [X] Empty Bands                                   :    55   450

I found it interesting that it was different in BSE.
In this calculation, I determined the unit-cell (6 atoms) as 1x1x2 as a super-cell (12 atoms).
It gave that error because I did not change the number of electrons in the old input.

Code: Select all

Nelectro= 50.00000               # Electrons number

I changed the number of electrons in the new input. Thus, I realized that the number of electrons in the input is important, or the necessity of producing new bse input in every calculation.

Code: Select all

Nelectro= 100.00000               # Electrons number

I attached input, output and report files. Is there any error in scf and nscf inputs?
The error of high results in EPS-Im and EPS-Re values continues.
I asked you about this issue before.

https://www.yambo-code.eu/forum/viewtopic.php?t=2624


I added Drude terms, or I increased the k-grid, or I changed the band ranges but still I couldn't solve this problem.
My last resort will be to try MPA with Yambo 5.2. I encountered the libxc-5.2.3 problem while compiling.

Best regards,
Zafer

[Daniele Varsano]

Dear Zafer,

Nelectro matters!! The number of electrons is read from QE, and it should not be touched/changed unless you want to dope your system.
The suggestion is to generate input files with command line and not using inputs produced for other systems, as Yambo reads relevant variables from databases when generating the inputs.

About the very high values on eps, it seems to me that it depends on the band structure, you can check dipole values and recognize if there are vertical transitions with nearly zero energy. In this case, your divergence in the eps is physical.

About the compilation, please provide more details, maybe posting in the compilation subforum.

Best,
Daniele