ypp, band interpolation, time-reversal symmetry

[Fabiof]

Sorry, i just edited the post to get the files attached.

[Daniele Varsano]

Dear Fabio,
let's try to fo step by step.
I do not know if your problem is related to TR symmetry anyway they should be removed first.
> ypp -n
uncomment the
#RmTimeRev
in the generated input file
and run ypp.

Now you should have a new directory called FixSymm.
Enter there and try to run a band structure interpolation, at KS, so without reading the QP database.
Let's' see if it works and next let's see the QP bands.

Let us know where the code fails.

Best,
Daniele

[Fabiof]

Ok, i runned ypp -n

Code: Select all

fixsyms                      # [R] Reduce Symmetries
RmAllSymm                   # Remove all symmetries
RmTimeRev                   # Remove Time Reversal

Now i create a directory FixSymm with a sub-directory SAVE (no errors here).

In FixSymm folder i try to run the following ypp file:

Code: Select all

bnds                         # [R] Bands
electrons                    # [R] Electrons (and holes)
cooIn= "rlu"                 # Points coordinates (in) cc/rlu/iku/alat
BANDS_steps=  30             # Number of divisions
% INTERPGrid
-1 |-1 |-1 |                             # Interpolation BZ Grid
%
ShellFac= 20.00000           # The bigger it is a higher number of shells is used
%QPkrange                    #  generalized Kpoint/Band indices
  1| 19|  2| 8|
%
%BKpts                       # K points of the bands circuit
0.00000 |0.00000 |0.00000 |
0.33333 |0.66667 |0.00000 |
0.00000 |0.50000 |0.00000 |
0.00000 |0.00000 |0.00000 |
%

and the result is just:

Code: Select all

<01s> [01] Game Over

I cant even create a ypp file in FixSymm directory.

[Daniele Varsano]

Dear Fabio,
remove the TR symmetry and not all the symmetries:

Code: Select all

fixsyms                      # [R] Reduce Symmetries
#RmAllSymm                   # Remove all symmetries
RmTimeRev                   # Remove Time Reversal

Now i create a directory FixSymm with a sub-directory SAVE (no errors here).

You should not create it, it should be created when you run ypp.

In Fixsymm, generate the band interpolation ypp.in and run it.
I suggest you to start from scratch, ie, remove the Fixsymm directory which is now present.

Best,
Daniele

[Fabiof]

Thanks for the answer,

You should not create it, it should be created when you run ypp.

Sorry, that's not what i wanted to say. The folder is created when run ypp. Anyway i started from scratch as you suggested.

Inside the folder FixSymm there is a new SAVE folder (with more files than the older one).

The problem is that inside the FixSymm folder i can't generate the band interpolation ypp file.

When i execute the command ypp -s b (inside the FixSymm folder), i get the following message

Code: Select all

<01s> [01] Game Over

Fabio

[Davide Sangalli]

Dear Fabio,
before running ypp to generate the input you need to run the yambo initialization for the new folder.

Then you will be able to use ypp.

D.

[Fabiof]

Dear Davide,

thanks, it works now!


Bu there is still a error in this material (Graphene).

When i run this ypp file (Skipping MPI initialization ):

Code: Select all

bnds                         # [R] Bands
electrons                    # [R] Electrons (and holes)
cooIn= "rlu"                 # Points coordinates (in) cc/rlu/iku/alat
BANDS_steps= 20              # Number of divisions
% INTERPGrid
-1 |-1 |-1 |                             # Interpolation BZ Grid
%
ShellFac= 20.00000           # The bigger it is a higher number of shells is used
%QPkrange                    #  generalized Kpoint/Band indices
  1| 31|  1| 10|
%
%BKpts                       # K points of the bands circuit
0.00000 |0.00000 |0.00000 |
0.33333 |0.66667 |0.00000 |
0.00000 |0.50000 |0.00000 |
0.00000 |0.00000 |0.00000 |
%

it gives the error " <01s> [INTERPOLATION] Number of shells: 627Segmentation fault"


If i run ypp without skipping MPI initialization i get the following error

Code: Select all

<01s> [INTERPOLATION] Number of shells:   627[search6:03819] *** Process received signal ***
[search6:03819] Signal: Segmentation fault (11)
[search6:03819] Signal code: Address not mapped (1)
[search6:03819] Failing at address: 0x2b100e393000
[search6:03819] [ 0] /lib64/libpthread.so.0[0x30a660f790]
[search6:03819] [ 1] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x48650b]
[search6:03819] [ 2] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x4810d8]
[search6:03819] [ 3] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x447a5c]
[search6:03819] [ 4] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x43289c]
[search6:03819] [ 5] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x412354]
[search6:03819] [ 6] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x40ddf1]
[search6:03819] [ 7] /lib64/libc.so.6(__libc_start_main+0xfd)[0x30a5a1ed5d]
[search6:03819] [ 8] /home/fabiof/bin/yambo-4.0.1-rev.89/bin/ypp[0x40d009]
[search6:03819] *** End of error message ***
Segmentation fault

This is not happening in isolator BN-h material where i can get the interpolated band (KS and QP) without any error.

Fabio

[Daniele Varsano]

Dear Fabio,
yes sorry I forgot to tell you to re-run the setup after removing the TR symmetry.
We will need to reproduce your error in order to spot the problem.
It is rather strange, consider that the system you are studying is the one we used in the example in the web page.

Best,
Daniele

[Fabiof]

Dear Daniele,

I changed the value ShellFac to 1.00000 and it works.

In BN system it works with ShellFac = 20.00000 but it doesn't work with for example ShellFac = 50.00000.

It seems the problem is related to the number of shells.

Fabio

[Davide Sangalli]

Dear Fabio,
I'v been able to reproduce your problem.

It is indeed due to a problem with the symmetry operation defined in the interpolation procedure.

To fix it open the file src/interpolate/bz_interp_setup.F and after the loop which end at line 72

Code: Select all

    !
 enddo

insert the line: if(abs(int_sop(1,1,1)+1._SP)<zero) int_sop=-int_sop

Code: Select all

    !
 enddo
 !
 if(abs(int_sop(1,1,1)+1._SP)<zero) int_sop=-int_sop

After that recompile ypp.

D.