ERROR: Incorrect normalization

[elham....@gmail.com]

Hello every body!

I am doing epsilon calculation for a Au slab. The calculation stops at the very beginning step with this error:

  

  ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

   Wavefunction is not normalized in file WFN

   abs(norm - 1) = 0.0000011

   iband =    31 ispin = 1 ik =    11

From proc      0: ERROR: Incorrect normalization.

///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

Does any body know what is the problem?

Thank you in advance,

Elham

[Zhenglu Li]

Hi Elham,

It means the wavefunction is not normalized.  It is the 11th kpoint.  You may check first whether this kpoint is near the BZ boundary, and if so, you probably need to increase your energy cutoff.

Best,

Zhenglu

--
You received this message because you are subscribed to the Google Groups "BerkeleyGW Help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to help+uns...@berkeleygw.org.
To view this discussion on the web visit https://groups.google.com/a/berkeleygw.org/d/msgid/help/181399b8-929f-4ccf-aa58-97df546d2fda%40berkeleygw.org.

[yaz...@gmail.com]

Hi Zhenglu

I have encountered a similar problem. However the k point in question is 0 0.1667 0, nowhere near the BZ boundary. The mf energy cutoff inQE is already 80 Ryd. The wavefunction is generated by parabands.x. 

Is this still a energy cutoff issue or there's something about parabands.x or something else?

Many Thanks

Yaze

p.s. here is the error detail:

 Wavefunction is not normalized in file WFN

   abs(norm - 1) = 0.9861495

   iband =   163 ispin = 1 ik =     2

On Friday, 3 April 2020 00:08:47 UTC+8, Zhenglu Li wrote:

Hi Elham,

It means the wavefunction is not normalized.  It is the 11th kpoint.  You may check first whether this kpoint is near the BZ boundary, and if so, you probably need to increase your energy cutoff.

Best,

Zhenglu

On Thu, Apr 2, 2020 at 12:12 AM <elham...@gmail.com> wrote:

Hello every body!

I am doing epsilon calculation for a Au slab. The calculation stops at the very beginning step with this error:

  

  ///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

   Wavefunction is not normalized in file WFN

   abs(norm - 1) = 0.0000011

   iband =    31 ispin = 1 ik =    11

From proc      0: ERROR: Incorrect normalization.

///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

Does any body know what is the problem?

Thank you in advance,

Elham

--
You received this message because you are subscribed to the Google Groups "BerkeleyGW Help" group.

To unsubscribe from this group and stop receiving emails from it, send an email to he...@berkeleygw.org.

[Zhenglu Li]

Hi Yaze,

Something is severely wrong in your wavefunction (not cutoff), as your norm is 1-0.9861495=0.01385, very far from 1.0.  You may check whether you do every step correctly. 

Best,

Zhenglu

[Yaze Wu]

Hi Zhenglu

 

I suspected the same thing and am re-running parabands.x using the output from the previous parabands.x run as input.

 

Going back to check the details, the kgrid input is correct and the initial bands calculation on QE  completed successfully. The only possible reason I can think of at the moment is that I used number_pools = 4 instead of 1 the last time round; however, as the k points are allocated to separate pools of cores for calculation, I don’t think using number_pools = 4 should be give a result any different from number_pools = 1. Nevertheless, I’m re-running using number_pools = 1 this time round.

 

Will update again if it’s resolved.

 

Best wishes

Yaze

[elham....@gmail.com]

Hello!

I also checked the kpoints used for the calculation there was no problem. Also, Increasing the cut off did not work. I could solve this problem with not using npool. I just used ndiag for the QE calculation.

Best Regars,

Elham

[Yaze Wu]

Hi Elham

 

Thanks for the info, this is definitely helpful!

Let me see if the pooling in parabands behaves the same and get back to you and Zhenglu when my calculations finish.

 

Best

Yaze

[Yaze Wu]

Dear Zhenglu and Elham

 

The error has disappeared after I used the wfn from number_pools = 1.

 

Hope this helps anybody bumping into this issue in the future.

 

Best wishes

[Yaze Wu]

[wxia...@gmail.com]

I experienced similar problem with finite momentum BSE calculations. However, using npools=1 doesn't work for my case. It is strange that the error didn't show up for standard kernel calculation, i.e. Q=0, but persist for finite Q. Anything wrong?

Best,

Xiaoming

[Jacopo Simoni]

I am having the same problem at finite Q, does anybody have some idea ?

[Lai MingRui]

Hi,

I am also facing the same issue with finite Q calculations. One way I found is just to use "dont_check_norm" in the input, however one must be sure that the wavefunction is correct.