[ERROR] ... in :[03] Transferred momenta grid

[devsharma]

hi 2 all,
I was trying to get familiar with the yambo, with using output produced by PWSCF. I used the version 4.0.2.. I tried on example 1 on Si. My input files are listed below. When, i run
/home/dev/yem/yambo-3.2.1-r.448/bin/yambo -F /home/dev/work/si/temp/si.save/01_init -J 01_init
i got error as
<---> [01] Job Setup
<---> [02] Input variables setup
<---> [02.01] Unit cells
<---> [02.02] Symmetries
<---> [02.03] RL shells
<---> Shells finder |####################| [100%] --(E) --(X)
<---> [02.04] K-grid lattice
<---> [02.05] Energies [ev] & Occupations
<---> [03] Transferred momenta grid

[ERROR] STOP signal received while in :[03] Transferred momenta grid
[ERROR][RL indx] 2 equivalent points in the rlu grid found
what does it mean ??? Please help.
Same error mssg when i used the next step
home/dev/yem/yambo-3.2.1-r.448/bin/yambo -F /home/dev/work/si/temp/si.save/in_02_RPA_no_LF -J 02_RPA_no_LF
I also tried to edit 01_init file from
setup # [R INI] Initialization
MaxGvecs= 800 RL # [INI] Max number of G-vectors planned to use
to
setup # [R INI] Initialization
MaxGvecs= 800 RL # [INI] Max number of G-vectors planned to use
QpntsRXd
1 | 10 | # [Xd] Transferred momenta

and on web it is given
It specifies the range of momenta at which the screening function is calculated format:
first-Qpoint | last-Qpoint|
but screening is not done in the calculations.

2). Secondly, i want to know that, is it mandatory to use the Norm conserving pseudo potentials in scf and bands calculations to get EELS and absorption.

Input file for scf

&control
calculation='scf'
restart_mode='from_scratch',
prefix='si',
pseudo_dir = '/home/dev/pseudo/',
outdir='/home/dev/work/si/temp',
wf_collect = .true.
/
&system
ibrav= 2, celldm(1) =10.20, nat= 2, ntyp= 1,
ecutwfc =18.0,nosym = .true., force_symmorphic = .true.,
/
&electrons
diagonalization='david'
mixing_mode = 'plain'
mixing_beta = 0.7
conv_thr = 1.0d-8

/
ATOMIC_SPECIES
Si 28.086 Si.pz-vbc.UPF
ATOMIC_POSITIONS
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS
10
0.1250000 0.1250000 0.1250000 1.00
/
0.3750000 0.3750000 0.6250000 3.00

Bands calculations
&control
calculation='bands'
restart_mode='from_scratch',
prefix='si',
pseudo_dir = '/home/dev/pseudo',
outdir='/home/dev/work/si/temp',
wf_collect = .true.
/
&system
ibrav= 2, celldm(1) =10.20, nat= 2, ntyp= 1,
ecutwfc =18.0, nbnd = 8,nosym = .true.,
/
&electrons
diagonalization='david',
/
ATOMIC_SPECIES
Si 28.086 Si.pz-vbc.UPF
ATOMIC_POSITIONS
Si 0.00 0.00 0.00
Si 0.25 0.25 0.25
K_POINTS
28
0.0 0.0 0.0 1.0
/
0.5 0.5 0.5 1.0

3) One mor thing, in scf run , mssg in terminal
[dev@dev si]$ sh submit.sh
Cannot match namelist object name force_symmorphic
namelist read: misplaced = sign
Cannot match namelist object name .true.
I also find force_symmorphic in input PW in doc help file in Espresso . I had not found this variable. Is this has been removed or changed ?????

then using /home/dev/yem/yambo-3.2.1-r.448/bin/p2y -S -N
i got Save directory with
[dev@dev SAVE]$ ls
db.gops wf_00004_00001 wf_00011_00001 wf_00018_00001 wf_00025_00001
s.db1 wf_00005_00001 wf_00012_00001 wf_00019_00001 wf_00026_00001
wf_00006_00001 wf_00013_00001 wf_00020_00001 wf_00027_00001
s.wf wf_00007_00001 wf_00014_00001 wf_00021_00001 wf_00028_00001
wf_00001_00001 wf_00008_00001 wf_00015_00001 wf_00022_00001
wf_00002_00001 wf_00009_00001 wf_00016_00001 wf_00023_00001
wf_00003_00001 wf_00010_00001 wf_00017_00001 wf_00024_00001
[dev@dev SAVE]$

Is the conversion from PWscf to Yambo is correct ????
Thanks
take care
Dev

[Conor Hogan]

Hi Dev,

[ERROR][RL indx] 2 equivalent points in the rlu grid found
what does it mean ??? Please help.

[rlu = reciprocal lattice unit]
It means that when Yambo was trying to work out the set of q-vectors (transferred momenta) from the input set of k-points, it found some points were overlapping. This probably means you did something weird in your input (see below)

I also tried to edit 01_init file from

Please dont edit the fields/keywords in the input file by hand, you will surely cause Yambo to do weird things.
Instead, use the appropriate command line option relevant for what you want to calculate to generate the appropriate input file. If you want to calculate the screening, then you should invoke a runlevel that will calculate the screening - i.e. not the initialization runlevel (-i), which just sets up some simple databases . See yambo -H for the allowed runlevels.

2). Secondly, i want to know that, is it mandatory to use the Norm conserving pseudo potentials in scf and bands calculations to get EELS and absorption.

Yes, completely mandatory. You might have to develop some monster pseudopotential for your Yttrium!

ecutwfc =18.0,nosym = .true., force_symmorphic = .true.,

Please see the thread "Si example from Quantum Espresso fails with Seg Fault" at viewtopic.php?f=14&t=104&start=0
to see how to handle the force_symmorphic stuff in scf and bands runs, you are doing it incorrectly.

K_POINTS
10
0.1250000 0.1250000 0.1250000 1.00
/
0.3750000 0.3750000 0.6250000 3.00

I cannot reproduce your error if you dont post the whole list of k-points....anyway, look at the linked forum before reposting.

[dev@dev si]$ sh submit.sh
Cannot match namelist object name force_symmorphic

I'm not sure that force_symmorphic is an allowed keyword in pw 4.0.x. Perhaps it is only in 4.1.x. Check Doc/INPUT_PW for your distribution. In any case you might not need it.

Is the conversion from PWscf to Yambo is correct ????

Yes - you split the wavefunctions using -S in directories for each k-point. This is not necessary if you have a small system like bulk silicon, but it works fine.
By the way I urge you to compile Yambo with the netCDF library linked - it makes things a lot easier in the long run, trust me.

Cheers
Conor

[andrea marini]

take care
Dev

Dev, please update your signature to include your complete affiliation.

Thanks

[devsharma]

Dear Conor Hogan,
Thanks for help. I studied the discussion at
viewtopic.php?f=14&t=104&start=0
and successfully repeated the Si calculations. Sir, but with the input file given by you on this link , i think we have to use nosym=.true., in nscf calculations as we are using the automatic option for k points, otherwise it was giving error i.e.

[dev@dev forum]$ sh submit.sh
Cannot match namelist object name force_symmorphic
namelist read: misplaced = sign
Cannot match namelist object name .true.&electronsdiago_full_acc
namelist read: misplaced = sign
Cannot match namelist object name .true.diago_thr_init
namelist read: misplaced = sign
Cannot match namelist object name 1.0d-6atomic_speciessi
Cannot match namelist object name 28.086
Cannot match namelist object name si.vbc.upfatomic_positionssi
Cannot match namelist object name 0.00
Cannot match namelist object name 0.00
Cannot match namelist object name 0.00si
Cannot match namelist object name 0.25
Cannot match namelist object name 0.25
Cannot match namelist object name 0.25k_points
Cannot match namelist object name {automatic}4
Cannot match namelist object name 4
Cannot match namelist object name 4
Cannot match namelist object name 1
Cannot match namelist object name 1
STOP 2
[dev@dev forum]$

and reading namelist control error in nscf.out. Sir well that is solved by setting nosym=.true..Thanks

Secondly, sir, regarding the norm conserving pseudo, i am not expert in creating pseudos .My group is totally experimental group . in this case, even if i designed a pseudo, nobody is here for checking it ,(except forums) weather it is right or not ???
Just, i want to ask, that if I convert Yttrium and Vanadium GGA.fhi pseudos (given at abinit site) into the UPF format for use in PWScf calculations, will it work or not ??? please give some guidance.

3) Surly , I will try to install yambo with netCDF library.

Thanks
with regards,
Dev Sharma
University of Delhi
India

[Conor Hogan]

Hi Dev,

Before you give Andrea a heart attack , please update your *signature* with your affiliation,
which you can do at: user control panel -> profile -> edit signature.

Cannot match namelist object name force_symmorphic

This is nothing to do with using nosym = .true. or not, as I said, this is because you are using elements of the input namelist which are not compatible with the version of pwscf that you are using! In any case, every version of pw seems to generate k-points a little differently, so you have to make a final judgement yourself on whether the k-point set is ok or not.

will it work or not

well, it will work in the sense that you will create a pseudopotential that will work in PWscf. However, unless you do plenty of checks, you might be generating something that will give you nonsense results. But this is outside the scope of this forum. Also note, as stated many times, that if you use p2y you will not have the information regarding the nonlocal part of the commutator for the dipole matrix elements ( [r,V_nl]), which may or may not be crucial for your case. If you have no idea what I am talking about, or no idea how large the nonlocal contribution will be for your Yttrium/Vanadium compound, I suggest you use Abinit for everything....

regards,
Conor

[andrea marini]

I just moved the last post by dev about NETCDF in the compilation sub-forum. We will continue the discussion there as it could be an helpful help for others.

Cheers