Beyond the Plasmon-Pole approximation in LiF inout database

[Daniele Varsano]

Dear Kopinjol,
Sorry to not have been clear enough. I was talking about experiment and literature,
but the point is the same also when comparing with other code working in different way (different basis set, eg plane waves vs real space).
Comparison are meaningful only if the calculations of both code are done at convergence and using
the same approximations. Now, Yambo parameter of the tutorial are not at convergence so this
is the first source of the discrepancies. Next, also very important, for reason of feasibility I told you in the last post (the calculation had to run in few minutes otherwise the students get bored, distracted and inactive), the BSE calculation
of the tutorial is not done following a previous GW calculation, but applying a scissor operator. This approximation is not justified for this
system, I would say totally unjustified for a molecule, here we are applying a 5.1eV scissor to the LDA gap, and I do not know know if the value
really correspond to the real QP gap. A different value of the scissor will perform a rigid shift of the spectrum. So I do not know now if in your
real-space calculation you are doing the same crude approximation.

Hope it helps,

Cheers,

Daniele

[kopinjol]

Hello Daniele,

Thank you so much for the reply. I understand now what you are talking about. I will properly converge the yambo calculations and then compare. I will let you know what i get.

Thanks again, I really appreciate the help.

I had another question, it is a technical issue. I was trying to run yambo for a single Ag ion. So I made a input file by

yambo -N -x -V RL 01HF.in.

Which looks like this:

HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
EXXRLvcs= 3 Ry # [XX] Exchange RL components
%QPkrange # [GW] QP generalized Kpoint/Band indices
1| 1| 1| 14|
%
%QPerange # [GW] QP generalized Kpoint/Energy indices
1| 1| 0.0|12.0|
%

Now as you see EXXRLvcs is assigned a value 3Ry. But when I run yambo by typing:

yambo -N -F 01HF.in -J HF_3Ry

The output file shows the value of EXXRLvcs as 0. At the same time it shows FFTGvecs as 7 RL, which I assume is the default value or the value read from the PWScf database.

Did you ever run into such a problem? I am attaching the output file along with the post.

Thanks again,
kopinjol

[Daniele Varsano]

Dear Kopjnol,
unfortunately I cannot read the file you attached.
Anyway:

The output file shows the value of EXXRLvcs as 0. At the same time it shows FFTGvecs as 7 RL, which I assume is the default value or the value read from the PWScf database.

something went wrong, and FFTGvecs as 7 RL cannot see the default as I imagine that the pwscf calculation used more than 7 G vectors. Try to type yambo -D in order to assure that the databases generated by p2y are ok and nothing strange happened during the p2y run.
What happen if you put the FFTGvecs and EXXRLvcs in RL units, ie the number of Gvector you want to use?

Cheers,
Daniele

[kopinjol]

Hello Daniele,

Thanks for the quick reply. Here's the output of yambo -D:

[RD./SAVE//ns.db1]------------------------------------------
Bands : 14
K-points : 1
G-vectors [RL space]: 338107
Components [wavefunctions]: 42211
Symmetries [spatial]: 48
Spinor components : 1
Spin polarizations : 1
Temperature [ev]: 0.000000
Electrons : 10.00000
WF G-vectors : 42211
Max atoms/species : 1
No. of atom species : 1
- S/N 007306 ---------------------------- v.03.02.05 r.*** -
[RD./SAVE//ns.wf]-------------------------------------------
Bands in each block : 14
Blocks : 1
- S/N 007306 ---------------------------- v.03.02.05 r.*** -
[RD./SAVE//ndb.kindx]---------------------------------------
Polarization last K : 1
QP states : 1 1
X grid is uniform :yes
BS scattering :no
- S/N 007306 ---------------------------- v.03.02.05 r.*** -


I specified the FFTGvecs and EXXRLvcs as 7 and 5 RL's respectively. Here is the output (the last 4 lines of the output file) of my yambo run:
.-Input file : 02HF.in
| HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
| FFTGvecs= 7 RL # [FFT] Plane-waves
| EXXRLvcs= 7 RL # [XX] Exchange RL components

So it read FFTGvecs correctly but didnt read EXXRLvcs. Instead it put EXXRLvcs = FFTGvecs.

Now I specified the FFTGvecs and EXXRLvcs as 7 and 5 Ry's respectively. Now I get the following in the output file (the last 4 lines of the report file):

.-Input file : 02HF.in
| HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
| FFTGvecs=0 Ry # [FFT] Plane-waves
| EXXRLvcs=0 Ry # [XX] Exchange RL components

I have attached both the output files. It seems yambo reads the FFTGvecs values when they are specified in RL's. It also seems that yambo is not reading the EXXRLvcs and just setting it equal to FFTGvecs. Could it be a technical program?

Thanks,
kopinjol

[Daniele Varsano]

Dear Kopjnol,
First of all,I imagine it is just simple test, but just in case, take in mind you are using 7 plane waves over more than 40000,
so your wave functions are not properly described, not normalized and so on, so you will have meaningless results.
You can see this from the report:

Code: Select all

[WF loader] Normalization (few states)  min/max  :0.541E-09 0.843

Next, yambo automatically put the value to 7 because it approximate to the superior values you indicate in order to
have closed shell, ie all the Gvec such that |G|^2 approach the value you indicate. This is done for symmetry
reason of you reciprocal space.
From your report (Section RL shells) you can see that the second shell contains up to 7 vector so, 5 it is not permitted.

Now, why it does not recognize the Ry I do not know, we will have a look, in the meanwhile
try to use mHa or work with RL. In the report files you will have the value corresponding to the first and last shell in RL number
and mHa (Section RL shells).

Note the exchange part of the self energy is the fastest one, so you can easily include all the G vectors for reasonable systems.
For the correlation part, it is not obvious and you have to perform convergences test.

Hope it helps,

Cheers,

Daniele

[kopinjol]

Hello Daniele,

Thanks for the reply. I will do as you said and let you know what I get. What you said makes sense. I guess I will run with RL and see how it works.

Thanks again,
kopinjol

[kopinjol]

Hello Danielle and everyone,

I am running into a weird problem with Yambo. It doesnt seem to read the inputs from the input file and only assigns the default values. Here is my input file:

optics # [R OPT] Optics
chi # [R CHI] Dyson equation for Chi.
alda_fxc # [R TDDFT] The ALDA TDDFT kernel
% QpntsRXd
1 | 1 | # [Xd] Transferred momenta
%
% BndsRnXd
1 | 14 | # [Xd] Polarization function bands
%
NGsBlkXd= 300 RL # [Xd] Response block size
% EnRngeXd
0.00000 | 20.00000 | eV # [Xd] Energy range
%
% DmRngeXd
0.10000 | 0.10000 | eV # [Xd] Damping range
%
ETStpsXd= 100 # [Xd] Total Energy steps
% LongDrXd
1.000000 | 0.000000 | 0.000000 | # [Xd] [cc] Electric Field
%
FxcGRLc= 300 RL # [TDDFT] XC-kernel RL size


When i run yambo with yambo -F 02ALDA.in -J 02ALDA , I get this at the end of the output (r-*) file, where it summarises the input:

.-Input file : 02ALDA.in
| optics # [R OPT] Optics
| chi # [R CHI] Dyson equation for Chi.
| alda_fxc # [R TDDFT] The ALDA TDDFT kernel
| FFTGvecs= 7 RL # [FFT] Plane-waves
| % QpntsRXd
| 1 | 1 | # [Xd] Transferred momenta
| %
| % BndsRnXd
| 1 | 14 | # [Xd] Polarization function bands
| %
| NGsBlkXd= 7 RL # [Xd] Response block size
| % EnRngeXd
| 0.00000 | 20.00000 | eV # [Xd] Energy range
| %
| % DmRngeXd
| 0.10000 | 0.10000 | eV # [Xd] Damping range
| %
| ETStpsXd= 100 # [Xd] Total Energy steps
| % LongDrXd
| 0.1000E-4 | 0.000 | 0.000 | # [Xd] [cc] Electric Field
| %
| FxcGRLc= 1 RL # [TDDFT] XC-kernel RL size

As you can see yambo sees NGsBlkXd= 7 and FxcGRLc= 1, when i had explicitly set their value to 300 each. I am attaching both input and output files as attachment. Does anyone know what may be happening?

Also I found a small bug in the one of yambo installation scripts (make-makefile.sh). For all C proprocessing I think we should use the "-traditional" flag, which was missing in the script. Without the "-traditional" flag the tabs and spaces are not copied to the Makefile and as such make fails with the error "missing tab" or something like that. I think it is also system dependent. Without the "-traditional" flag the code is not stable, meaning it compiles sometimes and shows error at other times arbitrarily.

Anyway Thanks in advance for all of your help.

Cheers,
kopinjol

[Daniele Varsano]

Dear Kopjinol,
it's hard to say what is happening form this information.
From the report I can see that yambo is doing nothing, the reported vaules are not
the default values but the values read form a previous calculation, you can see from the
report :

Code: Select all

[RD./02ALDA//ndb.dipoles]

RD means that yambo is reading a database, while WR when it writes a new one, so it seems
he is just repeating a calculation done before. Be sure you are
really running the new input and not a previous one. Try to copy your input 02ALDA.in in yambo.in
and to run yambo without the -F and -J options. What happen?

Cheers,

Daniele

[kopinjol]

Hello Daniele,

Thank you for pointing out that yambo is running some older calculation. Its weird but at least makes more sense than it not reading some of the input values!! I am going to remake the databases and try to run again.

Thank you so much, really appreciate your help.
kopinjol

[kopinjol]

Hi Daniele,

I figured out a way to run yambo so that it will recognize the input variables such that "MaxGvecs" . The way to make sure that yambo sees the input variable in the input file (01init.in) is to remake the database (p2y -N) before every init run. So every time I change the input value MafGvecs I remake the database and run yambo and yambo sees the input variable.

Having said that, I ran yambo for several values of MaxGvecs, everytime remaking the database, but I still found the same value for the homo-lumo gap. The various values of MaxGvecs are order of magnitude different and so finding the same value for the homo-lumo gap seems incorrect. Am I missing something? My system is a silver +1 ion.

Thanks,
kopinjol