ERROR STOP signal from G0W0 in silicon GW calculation

[okuno]

Dear Yambo Users and developlers.

I've just started Yambo with PWSCF interface.

I have calculate PWSCF scf and nscf calculation, and then run yambo GW calculation
but the calculation has stopped with the error message

[ERROR] STOP signal received while in [06.01] G0W0 : the P(lasmon) A(pproximation)
[ERROR] Incomplete and/or broken PPA/Static diel fun. databese.

I have calculate band number 200 in nscf and nosym=.true. option in PWSCF,
and use the yambo.in in the example.

Are there some inconsistency with the parameter in the yambo.in

QpntsRXp, or NGsBlkXP (where I used the sample yambo.in values.)

What number is reccomended for QpntsRXp, NGsBlkXP, for nbnd=200 with two si atom in the system?

Sincerely.

Yukihiro, Okuno.

[andrea marini]

Dear okuno,

Welcome on board! Please fill your signature with your complete name & affiliation. This is a simple request that we do to all members of this forum.

The reason why Yambo stops is that, as you can see from the r_setup, you have in total 256 Q-points. This is because you used no symmetries. However in the input file for the GW calculation you asked to calculate the screened interaction only on 29 q-points. As the self-energy is a sum over all the BZ you need to calculate W for ALL q-points.

My question is: why did you use no symmetries ? It would be enough to remove only the non-symmorphic ones. In this case you would have a total of 10 q-points saving a lot of calculation time.

What number is reccomended for QpntsRXp, NGsBlkXP, for nbnd=200 with two si atom in the system?

QpntsRXp mus always cover the whole range of q-points. NGsBlkXP can be estimated to be 1/10 to 1/5 of the wavefunction G-vectors. But you need to do very careful convergence tests in order to find the optimal parameters.

[okuno]

Dear Dr. Marini

Thank you for your reply. I'm glad to receive reply.
And sorry for not attach signature. I edited signature in user control panel.


>The reason why Yambo stops is that, as you can see from the r_setup, you have in total 256 Q-points. This is because you >used no symmetries. However in the input file for the GW calculation you asked to calculate the screened interaction only >on 29 q-points. As the self-energy is a sum over all the BZ you need to calculate W for ALL q-points.

Thank you. does this "Q-points" mean the IBZ Q-points in r_setup file , or K-points ?
( what is the difference of this Q-points and K-points ? )

>My question is: why did you use no symmetries ? It would be enough to remove only the non-symmorphic ones. In this case you >would have a total of 10 q-points saving a lot of calculation time.

In first, I use no symmetries in pwscf non-scf calculation, but when I run p2y, the error
has occurs like below,

<---> [01] P(W) 2 Y(ambo) Ver. 4.0
<---> DBs path set to .
<---> Index file set to data-file.xml
<---> Header/K-points/Energies...done
<---> Cell data...done
<---> Atomic data...done
<---> Symmetries...

[ERROR] STOP signal received while in :[01] P(W) 2 Y(ambo) Ver. 4.0
[ERROR] Non-symmorphic symmetry operations are not supported! Use nosym=.true.ed in PWSCF

The error says explicitly we should use nosym=.true. in PWSCF,
then, I change nosym=.true. in PWSCF, we could make database set for yambo, ./SAVE,.
Can I use non-scf calculation in PWSCF with nosysm=.false. ?

And in addition, I want t the NGsBlkXp's meaning ( which I thought the matrix size of \epsilon(G,G'),
G * G ? ) .

Sincerely.

Yukihiro Okuno

[Daniele Varsano]

Dear Okuno,
you can use symmetries, except the non-symmorphyc ones.
In the new release of PWSCF this issue can be solved using the keyword
force_symmorphic indicated in this post by Claudio.

viewtopic.php?f=7&t=25&p=253#p253

Thank you. does this "Q-points" mean the IBZ Q-points in r_setup file , or K-points ?
( what is the difference of this Q-points and K-points ? )

Yes they are the Q-points in the IBZ, obtained as K-K' (K-points) of your sampling of the BZ

And in addition, I want to know how to set the NGsBlkXp ( which I thought the matrix size of \epsilon(G,G'),
G * G ? ) .

Yes they are the size of the block of \epsilon(G,G'), and this is a variable that have to reach the convergences,
for instances in the linear response calculations this gives you the matrix size to invert in order to take into
account the local fields, it is a system dependent quantity, it depends on the disomogeneity of your system
and have to be checked.

Best regards,

Daniele

[claudio]

Dear okuno

there is a mistake in the message reported by p2y, in fact you should use force_symmorphic=.true. in PWSCF
and not nosym=.true. .
The error massage will be corrected in the new Yambo version

[okuno]

Dear Yambo Community members.

Thank you for your replies.
I must studied more about Yambo.

I had recalculated the si.nscf.in with force_symmorphic=.true. and nosym=.false.
Then p2y can produce the databaseset ./SAVE.

my procedure is
p2y
yambo -i
yambo
yambo -x -p p -g g -V 2

I could get the yambo.in like below.


xxvxc # [R XX] Hartree-Fock Self-energy and Vxc
ppa # [R Xp] Plasmon Pole Approximation
gw0 # [R GW] GoWo Quasiparticle energy levels
em1d # [R Xd] Dynamical Inverse Dielectric Matrix
FFTGvecs= 869 RL # [FFT] Plane-waves
EXXRLvcs= 4279 RL # [XX] Exchange RL components
XfnQPdb= "none" # [EXTQP Xd] Database
XfnQP_N= 1 # [EXTQP Xd] Interpolation neighbours
% XfnQP_E
0.000000 | 1.000000 | 1.000000 | # [EXTQP Xd] E parameters (c/v)
%
% QpntsRXp
1 | 85 | # [Xp] Transferred momenta
%
% BndsRnXp
1 | 200 | # [Xp] Polarization function bands
%
NGsBlkXp= 1 RL # [Xp] Response block size
CGrdSpXp= 100.0000 # [Xp] [o/o] Coarse grid controller
% EhEngyXp
-1.000000 |-1.000000 | eV # [Xp] Electron-hole energy range
%
% LongDrXp
1.000000 | 0.000000 | 0.000000 | # [Xp] [cc] Electric Field
%
PPAPntXp= 27.21138 eV # [Xp] PPA imaginary energy
GfnQPdb= "none" # [EXTQP G] Database
GfnQP_N= 1 # [EXTQP G] Interpolation neighbours
% GfnQP_E
0.000000 | 1.000000 | 1.000000 | # [EXTQP G] E parameters (c/v)
%
% GbndRnge
1 | 200 | # [GW] G[W] bands range
%
% GDmRnge
0.10000 | 0.10000 | eV # [GW] G_gw damping range
%
%QPkrange # [GW] QP generalized Kpoint/Band indices
1| 60| 1|200|
%
%QPerange # [GW] QP generalized Kpoint/Energy indices
1| 60| 0.0|-1.0|
%

I think in this yambo.in file that yambo generated, I think the important variable that I should edit are
QPkrange (or QPerange), NGsBlkXP, GbndRnge. BndsRnXp.
NGsBlkXp= 1 , is ok ? because if NGsBlkXp= 1 the matrix size of epsilon(G,G') =1 then G can take G=1 only.
And how should I take BndsRnXp and GbndRnge ?
Sincerely.
Yukihiro Okuno.

[claudio]

Dear Okuno

before running calculation you should understand the mining of each parameter.
About the parameters

NGsBlkXP http://www.yambo-code.org/doc/vars/var_NGsBlk.php
GbndRnge http://www.yambo-code.org/doc/vars/var_GbndRnge.php
BndsRnXp http://www.yambo-code.org/doc/vars/var_BndsRn.php

you should check the convergence of your results respect to these parameters.

While QPkrange (or QPerange) just referees to the Quasi-particles that you want to calculate.

To understand their meaning have a look here: http://www.yambo-code.org/doc/docs/doc_GW.php
or in the lecture notes on GW: http://www.yambo-code.org/lectures.php

Claudio

[okuno]

Dear Yambo developers and users.

Thank you for your suggestions.
I can calculate simple bulk silicon from PWSCF outputs.
I check the convergence of GbndRnge, NGsBlkXP el al.

By the way, I want to show the band structure or DOS of the results of the
GW calculation.

Are there any tools that plot the Band structure and DOS ?
I cannot find the information about the plotting the Band structure or DOS from the
results of Yambo.

Sincerely.

Yukihiro Okuno.

[claudio]

Dear Yukihiro Okuno

unfortunately there is not yet a tool to plot directly the quasi-particle band structure and DOS,
you have to copy the results form the o.qp file and do it by yourself.
Maybe in the next version we will add it

Claudio

[claudio]

Dear Yukihiro Okuno

in the last revision of yambo, that you can download using SVN

http://www.yambo-code.org/download.php

it has been implemented the possibility to plot the quasi-particle density of states with the command:

ypp -s s -V qp

Claudio