Yambo Community Forum

Posted: **Sun Aug 16, 2015 8:45 pm**

Dear Forum,

I am trying to perform GW calculations on a system with spin-orbit coupling.
I prepare the abinit KSS file, but ypp fails to reproduce the DFT bandstructure (The effect of SOC is missing there). According to abinit, time reversal symmetry is automatically switched off when nspinor=2 but a2y output says that the TR symmetry exists.

Code: Select all

____    ____    ___      .___  ___.  .______     ______
\   \  /   /   /   \     |   \/   |  |   _  \   /  __  \
 \   \/   /   /  ^  \    |  \  /  |  |  |_)  | |  |  |  |
  \_    _/   /  /_\  \   |  |\/|  |  |   _  <  |  |  |  |
    |  |    /  _____  \  |  |  |  |  |  |_)  | |  `--"  |
    |__|   /__/     \__\ |__|  |__|  |______/   \______/

 <---> A(binit) 2 Y(ambo)
 <---> Checking input file ...prkow_KSS
 <---> DBs path set to :.
 <---> KSS Header...abinit version 7.2.1
 <---> Pseudo Potential KB form factors...yes
 <---> Atom's informations...done
 <---> Symmetries...[SI no]...[I yes]...[-I no]...[TR yes]
 <---> RL vectors...done
 <---> Closing shells against inversion...done
 <---> :: ng_vec was increased to close the G-shells
 <---> Energies...done
 <28s> Report:
 <28s>  :: Electrons             : 80.00000
 <28s>  :: Temperature       [ev]: 0.000000
 <28s>  :: Lattice factors [a.u.]: 12.47496  12.47496  11.99515
 <28s>  :: K-points              : 256
 <28s>  :: Bands                 :  320
 <28s>  :: Spinor components     : 2
 <28s>  :: Spin polarizations    : 1
 <28s>  :: Spin orbit coupling   : yes
 <28s>  :: Symmetries   [spatial]: 1
 <28s>  ::                [T-rev]: yes
 <28s>  :: Max WF components     :  7042
 <28s>  :: RL vectors        (WF):  8008
 <28s>  :: RL vectors    (CHARGE):  8225
 <28s>  :: XC potential          : Slater exchange(X)+Perdew & Wang(C)
 <28s>  :: Atomic species        :  5
 <28s>  :: Max atoms/species     :  6
 <28s> == DB1 ...done ==
 <28s> == WFs + nlPP ...done ==

My abinit input parameters are:

Code: Select all


#Definition of the planewave basis set
ecut 500 eV
enunit 1

#Definition of the k-point grid
kptopt 1                # k points options
ngkpt 8 8 8 


#Definition of the SCF procedure
nstep 80          # Maximal number of SCF cycles
tolwfr 1.0d-18         # Will stop when this tolerance is achieved 
#Type of calculation
ixc 7
nspinor 2

# KSS inputs
kssform     3
nbandkss   320        # Number of bands in KSS file (-1 means the maximum possible)
nband      330        # Number of (occ and empty) bands to be computed
nbdbuf      10

Am I forgetting anything in abinit calculations?
I also tried to use ypp -n to remove the symmetry but the process stops at the WFs rotation step:

Code: Select all

 __    __ ______           ____     _____
/\ \  /\ \\  _  \  /"\_/`\/\  _`\ /\  __`\
\ `\`\\/"/ \ \L\ \/\      \ \ \L\ \ \ \/\ \
 `\ `\ /" \ \  __ \ \ \__\ \ \  _ <" \ \ \ \
   `\ \ \  \ \ \/\ \ \ \_/\ \ \ \L\ \ \ \_\ \
     \ \_\  \ \_\ \_\ \_\\ \_\ \____/\ \_____\
      \/_/   \/_/\/_/\/_/ \/_/\/___/  \/_____/

 <---> [01] Y(ambo) P(ost)/(re) P(rocessor)
 <---> [02] Core DB
 <---> :: Electrons             : 80.00000
 <---> :: Temperature       [ev]: 0.000000
 <---> :: Lattice factors [a.u.]: 12.47496  12.47496  11.99515
 <---> :: K points              : 256
 <---> :: Bands                 :  320
 <---> :: Symmetries            : 2
 <---> :: RL vectors            :  8225
 <---> [03] K-point grid
 <---> :: Q-points   (IBZ): 260
 <---> :: X K-points (IBZ): 256
 <---> [04] CORE Variables Setup
 <---> [04.01] Unit cells
 <---> [04.02] Symmetries
 <---> [04.03] RL shells
 <---> [04.04] K-grid lattice
 <---> [04.05] Energies [ev] & Occupations
 <---> Setting I/O path to "FixSymm"
 <---> [M  4.782 Gb] Alloc WF ( 4.768)
 <02s> [FFT-Magn] Mesh size: 25  25  25
 <30s> [WF-Magn loader] Wfs (re)loading |####################| [100%] 28s(E) 28s(X)
 <33s> Total magnetization (x,y,z) [Bohr_magneton] : 0.00      0.00      0.00
 <33s> [M  0.014 Gb] Free WF ( 4.768)
 <34s> [05] == Remove Time Reversal ==
 <34s> [05.01] Symmetries
 <34s> 1 spatial symmetries reduced/increased to ...1
 <34s> [Symm] Spatial simmetries now are:
 <34s> [S1] 1.0000  0.0000  0.0000  0.0000  1.0000  0.0000  0.0000  0.0000  1.0000
 <34s> [SYMs] Group table built correctly
 <34s> [05.02] K-points
 <34s> 256 k-points are now expanded/reduced to ... 512
 <34s> [05.03] Energies and WFs shells
 <34s> :: I/O...done
 <34s> [05.04] Wavefunctions
 <34s> :: Copying existing database ...done
 <34s> WFs rotation |                    | [000%] --(E) --(X)
forrtl: severe (24): end-of-file during read, unit 0, file /dev/pts/24
Image              PC                Routine            Line        Source             
ypp                00000000006B627E  Unknown               Unknown  Unknown
ypp                00000000006B4D16  Unknown               Unknown  Unknown
ypp                000000000066F802  Unknown               Unknown  Unknown
ypp                000000000062C45B  Unknown               Unknown  Unknown
ypp                000000000062B9C2  Unknown               Unknown  Unknown
ypp                00000000006476BB  Unknown               Unknown  Unknown
ypp                00000000004E2BB8  Unknown               Unknown  Unknown
ypp                00000000004D50FF  Unknown               Unknown  Unknown
ypp                00000000004A456B  Unknown               Unknown  Unknown
ypp                0000000000451E8B  Unknown               Unknown  Unknown
ypp                00000000004440A9  Unknown               Unknown  Unknown
ypp                0000000000407EC9  Unknown               Unknown  Unknown
ypp                0000000000405581  Unknown               Unknown  Unknown
libc.so.6          00007FA7B1B56D5D  Unknown               Unknown  Unknown
ypp                0000000000404799  Unknown               Unknown  Unknown

Does anyone know how to solve this problem?

Thanks in advance
Cheers,
Javad

Posted: **Sun Aug 16, 2015 8:53 pm**

Sorry, I forgot to mention that I am using yambo-3.4.1.
Cheers,
Javad

Posted: **Mon Aug 17, 2015 5:31 pm**

Dear Javad,
why do you say: "According to abinit, time reversal symmetry is automatically switched off when nspinor=2" ?
Indeed spin-orbit coupling breaks time reversal only if the system is magnetic.

I tried a simple run with a system with SOC and no magnetization.
abinit uses the time reversal symmetry and yambo recognizes it correctly. If I then remove the time_reversal the number of k-points is increased.

Instead the "seg fault" you experience with ypp should not happen. I'll have a look to that.
But again it is not related to the presence of the time reversal symmetry.

Kind regards,
Davide

Posted: **Mon Aug 17, 2015 5:40 pm**

Dear Davide,

Thank you for your prompt response. In abinit description for "istwfk" it says "For spin-orbit calculations (nspinor=2), istwfk is also forced to be 1, for all k points."
http://www.abinit.org/doc/helpfiles/for ... tml#istwfk

Isn't that the parameter that we should set?

Cheers,
Javad

Posted: **Mon Aug 17, 2015 6:24 pm**

Dear Javad

"istwfk" only regards how the wavefunctions are stored, not if the calculation uses the TR symmetry.
In fact I am always using the value 1 for any calculation. In fact I do not see an automatic way to remove the RT from the calculations in abinit. The only option seems to set "symrel" by hand excluding TR symmetries.

Best,
Myrta

Posted: **Mon Aug 17, 2015 7:51 pm**

Thank you Myrta for your answer,

So Something like symrel 1 0 0 0 1 0 0 0 1 and nsym 1 should do the job? Isn't it the default symrel?

Best Regards,
Javad

Posted: **Mon Aug 17, 2015 8:31 pm**

Hi Javad

In abint by default symrel is calculated automatically (nsym=0).
From the log of a2y that you posted it seems that what you proposed should do the work.

Best,
m

Posted: **Mon Aug 17, 2015 9:34 pm**

Hello again Myrta and thanks.

I already tried it and the a2y log still shows TR symmetry! Can you think of any reason for that?

Best,
Javad

Posted: **Tue Aug 18, 2015 11:50 am**

Dear Javad,
abinit does not write anywhere if the time-reversal symmetry is used or not inside the KSS file.

Thus, when a2y reads the KSS, it always assumes that the t-rev is used unless there is a magnetic system with the SOC.
If you want to convert with a2y a KSS file generated without time-reversal in abinit you have to manually modify the code in the interface to switch off the time reversal ...

However I'm not sure that setting nsym=1 in the abinit input will generate a KSS without the time-reversal.
I think nsym refers to the number of spatial simmetries only. How many kpts do you have after your DFT run 256 or 512 ?

Another way to remove the t-rev is to use ypp.
A coupe of questions to fix the seg-fault you experienced with ypp:
1) You are using yambo 3-4-1, but which revision ? Did you download the tar.gz file from the website or the source from svn ?
In case you are using the tar.gz can you try with the svn version ?
2) If you experience the seg-fault with the svn version, can you provide also the pseudo-potentials and the .files you use in input for abinit.

Best,
D.

Posted: **Tue Aug 18, 2015 8:22 pm**

Dear Davide,

For the KSS generation I used kptopt=3 which does not take any symmetry into account for k-points. Now a2y sees the full 512 k-points. However when I run yambo -N I see this output:

Code: Select all

 ___ __  _____  __ __  _____   _____
|   Y  ||  _  ||  Y  ||  _  \ |  _  |
|   |  ||. |  ||.    ||. |  / |. |  |
 \_  _/ |. _  ||.\_/ ||. _  \ |. |  |
  |: |  |: |  ||: |  ||: |   \|: |  |
  |::|  |:.|:.||:.|:.||::.   /|::.  |
  `--"  `-- --"`-- --"`-----" `-----"

 <---> [01] Files & I/O Directories
 <---> [02] CORE 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
 <01s> [03] Transferred momenta grid
 <01s> [WARNING][RL indx] 2 equivalent points in the rlu grid found
 <01s> [RL indx] X grid is not uniform.  Gamma point only.
 <01s> [04] Game Over & Game summary

and the outcome band structure is still wrong.

The yambo version is yambo-3.4.1-rev61.tgz, I couldn't use the svn version yet, but here you can find the input files for KSS and the psp files (was too big to be attached here) :
https://drive.google.com/file/d/0B55KC1 ... sp=sharing

Best Regards,
Javad

Yambo Community Forum

Fail to reproduce DFT bandstructure

Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure

Re: Fail to reproduce DFT bandstructure