NaN in PPA calculation

[mferri]

Dear developers,

I'm using Yambo 3.4.1 to perform different kinds of QP corrections on my germanane 2D system.
If I use the COHSEX approximation, everything is ok and I get the QP correction from o.qp file.
A problem arises when I use the PPA approximation (-p p - g n): the values of (E-Eo) and Sc(Eo) are NaN for each K-point and each band.

# K-point Band Eo E-Eo Sc(Eo)
#
1.00000 1.00000 -10.98441 NaN NaN
1.000000 2.000000 -9.210250 NaN NaN
1.000000 3.000000 -7.527469 NaN NaN

Looking at the report file, I noticed something strange in the Newton solver step.

[06] Dyson equation: Newton solver
==================================
[Newton] SC iterations :0

While in the COHSEX calculation

[06] Dyson equation: Newton solver
==================================

[Newton] Sc step [ev]: 0.10000
[Newton] Sc steps : 1
[Newton] SC iterations :0

It seems like it didn't solve the Dyson equation, and in the next step i read

QP [eV] @ K [1] (iku): 0.03571 0.06250 0.00000
B=1 Eo=-10.98 E= NaN E-Eo= NaN Z= NaN nlXC=-17.81259 lXC=-10.55430 So= NaN
B=2 Eo= -9.21 E= NaN E-Eo= NaN Z= NaN nlXC=-17.32113 lXC=-10.81398 So= NaN

I tried to change the solver from Newton to secant but the results are the same.
Could you please help me?

Best,

Matteo

[mferri]

I forgot to upload the attachment with the DFT and Yambo files,

Best,

Matteo

[Daniele Varsano]

Dear Matteo,
could you please post the complete input/output/report files?

Best,
Daniele

[mferri]

Dear Daniele,
sorry i forgot to attach the file in the first post, i re-upload it here.

Best,
Matteo

[Daniele Varsano]

Dear Matteo,
it is not straightforward to understand what is happening, most probably we would need to reproduce your error.
What I can suggest you is the following:
1) Update to the last release of the code (4.0.2) and see if the problem persists. Please note that some variable changed.
2) Not related with the error but in general, at least you really want them from some reason, it is not needed to calculate qp corrections for all the bands:

Code: Select all

  1| 28|  1| 55|

Are you interested in all the bands? usually one is interested to the bands around your gap/Fermi energy. You can limit to that bands saving a lot of time.

Best,
Daniele

[mferri]

Dear Daniele,

Thank you for the reply. I will ask the technician if it's possible to update the version of Yambo on our cluster.
About the point 2) you're right, I don't really need all the QP corrections, I will reduce this value. Thanks for the advice.

Best,

Matteo

[Davide Sangalli]

Dear Matteo,
I would need the pseudo of Ge to try to reproduce your problem.
I did not find it online.

Best,
Davide

P.S.: nice to know there is still someone in the MDM lab using yambo

[mferri]

Dear Davide,

I sent you an e-mail with the pseudo you were looking for.

Best,

Matteo

P.S. : just me and Alberto are using Yambo here, the last theorists remaining

[Davide Sangalli]

Ciao Matteo,
I tryed to reproduce the error but the runs are quite long (on my desktop the PPA takes more than 1 day).

Could you check if you are able to reproduce the same problem (i.e. NaN in the output) with a minimal set of parameters.
I.e. less bands, kpotints, G-vectors, in the yambo input.
You can start from the input file below.

Cheers,
Davide

Code: Select all

em1d                         # [R Xd] Dynamical Inverse Dielectric Matrix
ppa                          # [R Xp] Plasmon Pole Approximation
HF_and_locXC                 # [R XX] Hartree-Fock Self-energy and Vxc
gw0                          # [R GW] GoWo Quasiparticle energy levels
BoseTemp= 0.000000     eV    # Bosonic Temperature
FFTGvecs= 10063        RL    # [FFT] Plane-waves
EXXRLvcs=  1000       RL    # [XX] Exchange RL components
Chimod= ""                   # [X] IP/Hartree/ALDA/LRC/BSfxc
% BndsRnXp
  1 |  20 |                 # [Xp] Polarization function bands
%
NGsBlkXp= 1            Ry    # [Xp] Response block size
% LongDrXp
0.1000E-4 | 0.000    | 0.000    |        # [Xp] [cc] Electric Field
%
PPAPntXp= 27.21138     eV    # [Xp] PPA imaginary energy
% GbndRnge
   1 |  20 |                 # [GW] G[W] bands range
%
GDamping=  0.10000     eV    # [GW] G[W] damping
dScStep=  0.10000      eV    # [GW] Energy step to evalute Z factors
DysSolver= "n"               # [GW] Dyson Equation solver (`n`,`s`,`g`)
%QPkrange                    # [GW] QP generalized Kpoint/Band indices
  1|  1|  1| 20|
%

[mferri]

Dear Davide,
I reproduced the calculation you asked for, but the result is again NaN.

I would like to ask you what is the meaning of this variable

BoseTemp= 0.000000 eV # Bosonic Temperature

I noticed that the default value was -1.000 eV. Is it a conventional value?
Thanks,

Best,
Matteo