X grid is not uniform

[jjmm1974]

Dear all

I am trying to use yambo for a transparent conducting oxide. The DFT runs were made by Quantum Espresso 5.1.2, both the scf and nscf calculations with the option "force_symmorphic=.true." to avoid the use of non-integer translations and with the "automatic" option to sample the first Brillouin zone. I think that this should give a uniform grid of k points. However, upon yambo initialization, I get


<---> [01] CPU structure, Files & I/O Directories
<---> CPU-Threads:1(CPU)-8(threads)
<---> [02] CORE Variables Setup
<---> [02.01] Unit cells
<---> [02.02] Symmetries
<---> [02.03] RL shells
<---> Shells finder | | [000%] --(E) --( <05s>
Shells finder |############################ | [071%] 05s(E) 06 <10s>
Shells finder |########################################| [100%] 09s(E) 09s(X)
<10s> [M 0.026 Gb] Alloc RL_Gshells RL_Eshells ( 0.016)
<10s> [02.04] K-grid lattice
<10s> [02.05] Energies [ev] & Occupations
<10s> [03] Transferred momenta grid
<10s> [RL indx] X grid is not uniform. Gamma point only.
<10s> [04] External corrections
<10s> [05] Game Over & Game summary


I have checked the set of kpoints, and it seems fine to me. It is evident that I am missing something, but I am not able to figure it out. Does anyone have any hints about what may be going on here?

Thanks in advance.

Juanjo

P.S. I am attaching the input and output files for the scf and nscf calculations, as well as the PP I used.

Juan J. Meléndez
Department of Physics
University of Extremadura
Avda. de Elvas, s/n Badajoz (Spain)

[martinspenke]

Dear Juan,

I just performed a very quick test on your input files only for scf and it works for me (attached).
The grid is uniform.
In your nscf input file i see calculation='bands', it should be calculation='nscf'.

Bests
Martin

[jjmm1974]

Dear Martin

Thanks for your response. Indeed, I rerun the calculations from the scratch and I got the (expected) uniform k-sampling. It seems that there were some corrupted file in the temporary folder or something strange. Anyway, I can go on now.

Thank you again.

Juanjo

[andrea.ferretti]

Dear Juan,

I've also noticed that in your QE input you have set conv_thr=1.d-4,
which is totally fine for the sake of reducing the execution time when running a test, while it looks somehow too coarse for a production calculations (I would set 1.0d-8, at least)

take care
Andrea

[jjmm1974]

Hi Andrea

Thanks for your suggestion. I have a relatively large system (40 atoms in the primitive unit cell), and I thought that conv_thr=1.d-4 (which, I think, is an extensive quantity) would suffice to get adequate KS eigenvalues. I will decrease the parameter.

Take care

Juanjo