error running yambo

[samaneh]

Hi there,
I run yambo -g n -p p
and then change the parameters Gbndrnge, Bndsrnxp, NgsBlkxp
while I have the error(which is included in the attached file) in running yambo.
would you please guide me about the error??
Thanks
S. Ataei, University of Tehran, Iran

[Daniele Varsano]

Dear Samaneh,
the run you posted relative to the input file you included it looks it finished correctly.
Anyway, you should add in the input file:
Chimod= "Hartree",
or prepare the input using:

Code: Select all

yambo -p p -k hartree -g n

If the problem persists please post the failed input file together with the report and standard output.

Best,
Daniele

[samaneh]

Hi Daniele,
I run again yambo -g n -p p and add the Chimod= "Hartree" in input file
but it has error like the previous error!!!
I attached the needed files
thanks

[Daniele Varsano]

Dear Samaneh,
I think the problem is realted to the fact you ask for 50 bands in G summation:

Code: Select all

% GbndRnge
   1 |  50 |                 # [GW] G[W] bands range
%

...but you calculated only 32 in your DFT calculation:

Code: Select all

 [RD./SAVE//s.db1]-------------------------------------------
  Bands                           :  32
  K-points                        : 30

do the run will go out-of-bounds. Reduce the GbndRnge, or more appropriately, calculate more bands in your nscf run, as most probably you will need them considering you have 24 occupied bands.

Best,
Daniele

[samaneh]

Dear Daniele,
1-How can I change these parameters (Gbndrnge, Rnbndxp) in the input file
of yambo????
these values should be optimized!!!!
whenever I change the values in input file it has error like the previous error
which I attached!!!
2-I run with more band in nscf but actually
when I plot the band structure (using espresso plotband.x)
it has a wide band gap (around 8 eV) which is far from 3.6 eV(real value)
3-and all the band structures with different k points or different NGs are the same!!!!
I attached bandplot
would you please help me about these problems

[Daniele Varsano]

Dear Samaneh,
1) These value have to be converged: of course you cannot rise these values if you did not have calculate that number of bands before!!!
2) In the run you posted before you have:

Code: Select all

 Indirect Gaps      [ev]: 2.181099  2.930857
  Direct Gaps        [ev]: 2.453299  4.090193

so I would say quite consistent considering that PBE gap is not the QP gap. If you get now 8eV, either there is a problem with your nscf input file, either in the ground state density you calculated with the scf run, either in the use of plotband utility, or other problems I cannot know.
For help on this you should ask in the qe mailing list.
3) Most probably your calculation is converged wr kpoints and NGs and the problem you are facing is related to other reasons.
I do not know what happened to your calculation , in line of principle you should just set the nbnd variable in the nscf runlevel to a bigger value in the &system namelist, next you can raise the number of bands and in Xp up to that value.

Best,
Daniele

[samaneh]

Dear Daniele,
Would you please tell me how can I plot the
band structure using o.band-sp1 out put file of ypp,
I mean what is the plotband utility to perform it?
Best regards

[Daniele Varsano]

Dear Samaneh,
please fill you signature with your affiliation.
You can plot the ouptut file using gnuplot.
In the header it is specified what is each column:

Code: Select all

#   kx         ky         kz         b1         b2    b3...

so launching gnuplot and using:

p 'o.bands-sp1' u 0:4 notitle w l lw 1.5 lc 1,'' u 0:5 notitle w l lw 1.5 lc 1,....

you can plot the band structure. If you are not familiar with gnuplot you can use xmgrace or any other software to plot it.

Best,
Daniele