# DFPT\_CGWF WARNING : d2edt2 is zero, skipping update

**URL:** <https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369>\
**Category:** Ground state\
**Tags:** optic\
**Created:** [July 4, 2023, 7:42am UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369 "2023-07-04T07:42:10Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![jiang](https://discourse.abinit.org/letter_avatar_proxy/v4/letter/j/3e96dc/32.png) [@jiang](https://discourse.abinit.org/u/jiang)\
**Post date:** [July 4, 2023, 7:42am UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369/1 "2023-07-04T07:42:10Z")

</div>

Hey everyone, I’m getting a warning when calculating the nonlinear optic susceptibility that wasn’t mentioned in the tutorial（[https://docs.abinit.org/tutorial/nlo/#1-ground-state-properties-of-alp-and-general-parameters](https://docs.abinit.org/tutorial/nlo/#1-ground-state-properties-of-alp-and-general-parameters)）. Have any of you encountered this issue before? And how can I get rid of this warning?Thanks！

 ![image](https://discourse.abinit.org/uploads/default/original/1X/48148751529855a98db56889ce15e7b44e7b9da4.png)

---

<div class="post-metadata">

**Author:** ![bguster](https://discourse.abinit.org/letter_avatar_proxy/v4/letter/b/ac8455/32.png) [@bguster](https://discourse.abinit.org/u/bguster)\
**Post date:** [July 6, 2023, 3:11pm UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369/2 "2023-07-06T15:11:58Z")

</div>

Dear Jiang,

Essentially WARNING messages in Abinit should be **almost** harmless to the physical properties you are calculating, otherwise you have the ERROR messages that put a hard stop to your calculation when an incongruent set of parameters is provided.

In your case, if you want to make sure about what’s going on, you can increase the amount of printed information with prtvol (a value of 11 or higher should do the job). You will be able to see for which band the d2edt2 value is zero.  
My suspicion is data the number of bands or nbdbuf have been changed from one dataset to another.  
Let me know!  
Bogdan

---

<div class="post-metadata">

**Author:** ![jiang](https://discourse.abinit.org/letter_avatar_proxy/v4/letter/j/3e96dc/32.png) [@jiang](https://discourse.abinit.org/u/jiang)\
**Post date:** [July 7, 2023, 4:12am UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369/3 "2023-07-07T04:12:27Z")

</div>

Dear Bogdan

You are right. I tested nbdbuf and the warning disappeared when nbdbuf was set to 0. Whether the warning is present or not, the output results are the same, and the war  
ning does not affect the final result. Unfortunately, even when I increased the accuracy to 30 Ha, 16_16_16 k points, the nonlinear optical susceptibility still had a significant difference from the literature([https://journals.aps.org/prb/pdf/10.1103/PhysRevB.71.125107](https://journals.aps.org/prb/pdf/10.1103/PhysRevB.71.125107)). I have attached the input file for AlAs, and I would be grateful if someone could tell me where the problem is.  
[tnlo\_2.abi](https://discourse.abinit.org/uploads/short-url/s5gYPXONMOv4T0L7xhPKbuvl47p.abi) (5.7 KB)  
[tnlo\_2.abo](https://discourse.abinit.org/uploads/short-url/1dzptea6SV1PYV68KJnqleNr8pt.abo) (179.2 KB)  
[tnlo\_4.abi](https://discourse.abinit.org/uploads/short-url/xP9EIxEoOER9E02hN8QIVaxT9Co.abi) (1.3 KB)  
[tnlo\_4.abo](https://discourse.abinit.org/uploads/short-url/dyDBIDNTjQKgQSsbJQokeDWdWRG.abo) (17.0 KB)

---

<div class="post-metadata">

**Author:** ![ebousquet](https://discourse.abinit.org/letter_avatar_proxy/v4/letter/e/67e7ee/32.png) [@ebousquet](https://discourse.abinit.org/u/ebousquet)\
**Post date:** [July 7, 2023, 6:50am UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369/4 "2023-07-07T06:50:24Z")

</div>

Dear Jiang,  
You should (always) test the convergence w.r.t. kpoints and ecut, here kpoints should be quite enough (it might be more than necessary) but ecut=30Ha might be a low end value for norm-conserving pseudopotentials. Non-linear responses are quite “delicate” to obtain and you have to be sure that you have enough plane waves and k-points.  
In any case, convergence test should always be done for whatever property you calculate.  
Best wishes,  
Eric

---

<div class="post-metadata">

**Author:** ![jiang](https://discourse.abinit.org/letter_avatar_proxy/v4/letter/j/3e96dc/32.png) [@jiang](https://discourse.abinit.org/u/jiang)\
**Post date:** [July 29, 2023, 10:49am UTC](https://discourse.abinit.org/t/dfpt-cgwf-warning-d2edt2-is-zero-skipping-update/3369/5 "2023-07-29T10:49:15Z")

</div>

Dear Everyone,  
I’m sorry.My previous statement was incorrect. After tests, it is necessary to strictly limit nbdbuf to 0 in the calculation. Unfortunately, even when I set the ecut to 70 Ha and the ngkpt to 16_16_16, I still cannot obtain reasonable results.I hope someone can give me advice.Thanks!  
Best wishes,  
Jiang  
[tnlo\_4.abi](https://discourse.abinit.org/uploads/short-url/xP9EIxEoOER9E02hN8QIVaxT9Co.abi) (1.3 KB)  
[tnlo\_4.abo](https://discourse.abinit.org/uploads/short-url/4j5myzXiQocEbt2GFJz7CRif8eB.abo) (17.0 KB)  
[tnlo\_6.abi](https://discourse.abinit.org/uploads/short-url/4J2kr5C5lWyAVsiJu9Bj8fX8uDc.abi) (7.3 KB)  
[tnlo\_6.abo](https://discourse.abinit.org/uploads/short-url/omMFPuxjp0ltW0carnxptikKXfm.abo) (256.9 KB)