[CP2K-user] [CP2K:21038] Re: How to improve energy convergence?

[Frederick Stein]

Dear Hana,
Maybe some of the other users/developers are of better help. From the 
manual, you will find the following parameters:
- REL_CUTOFF (higher values increase the accuracy)
- parameters in the DFT/QS section
- parameters in the SUBSYS/KIND section (especially  LEBEDEV_GRID and 
RADIAL_GRID for the integration grids)

In the user group, there is the following interesting conversations:
https://groups.google.com/g/cp2k/c/FUvjXq9RvWY

Best,
Frederick

Hana schrieb am Donnerstag, 9. Januar 2025 um 21:10:04 UTC+1:

> Dear Frederick,
>
> Thanks for you response. I checked the convergence of EPS_DEFAULT and it 
> does converge under the defined threshold (1.0E-12). Also removing  
> EPS_PGF_ORB didn't help much. Do you have other suggestions? Thanks!
>
> On Tuesday, January 7, 2025 at 2:22:11 AM UTC-6 Frederick Stein wrote:
>
>> Dear Hana,
>> Can you check  the convergence of EPS_DEFAULT (and remove EPS_PGF_ORB)? 
>> The error in total energies is at the order of SQRT(EPS_DEFAULT) which is 
>> the order of magnitude of your deviations.
>> Best,
>> Frederick
>>
>> Hana schrieb am Dienstag, 7. Januar 2025 um 05:14:19 UTC+1:
>>
>>> Hello,
>>>
>>> I am working on a system consisting of EMIM and NTF2 ions. I am trying 
>>> to perform convergence test on the total energy of the system. Using the 
>>> GAPW method, the energy converges only up to the fourth decimal and this 
>>> convergence happens around a cutoff of 800 Ry. I had read in the forum that 
>>> GAPW is supposed to lower the required cutoff, but there seems to be an 
>>> issue in my case. I would appreciate it if you could take a look and let me 
>>> know about any ways to improve this convergence.
>>>
>>> 800     -311.670573521991457
>>> 700     -311.670595504244886
>>> 600     -311.670574077220238
>>> 500     -311.670578743875410
>>> 440     -311.670607726234607
>>> 420     -311.670640376258746
>>> 400     -311.670594145176324
>>> 380     -311.670596179369568
>>> 360     -311.670587449418463
>>>
>>> Thank you
>>>
>>>
>>>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/d5dbbe7b-8dbe-4540-ba93-f9b81a9b7637n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250113/ae368342/attachment.htm>