[CP2K-user] [CP2K:20505] Re: Extremely High Temperature and Pressure Increase using NPT

[Frederick Stein]

There is a howto on optimizing the CUTOFF in the manual 
(https://manual.cp2k.org/trunk/methods/dft/cutoff.html). In your case, you 
should check the forces and stress tensor (especially the diagonal 
elements) as well (signed values and sum of atomic forces) by turning on 
the respective PRINT section 
(https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/PRINT/FORCES.html, 
https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/PRINT/STRESS_TENSOR.html). 
Beware that stress tensors are generally not very accurate because with 
changes of the box dimensions, the integration grids will also change.

Michela Benazzi schrieb am Freitag, 26. Juli 2024 um 17:48:41 UTC+2:

> Hi Frederick, thank you very much for sharing this!
>
> The spike occurs as soon as the simulation starts, hence why I shared the 
> first couple frames. Would you still like me to include more?
>
> - CP2K is not printing any warnings.
> - Do you have any suggestions on what to look for when checking for 
> accuracy? I have changed CUTOFF in the past.
> - I could probably afford a larger basis set. The current one I am using 
> was the same used in prior research in my lab, where admittedly they did 
> not do much benchmarking! I can read more about zeta quality.
> - I will also fix smearing. 
>
> Sincerely thank you so much!! This gave me a good starting point and 
> different opinions. Best,
>
> Michela
>
> On Friday, July 26, 2024 at 10:43:53 AM UTC-4 Frederick Stein wrote:
>
>> Dear Michela,
>> Your output is not really helpful to me, it would be better to have a few 
>> frames around the spike. Does CP2K print any warnings? Spikes within MDs 
>> may hint to too inaccurate setups (or just bad luck). My suggestions are:
>> - did you check the accuracy of your setup (especially CUTOFF, 
>> EPS_DEFAULT)
>> - are you sure that a single zeta basis set provides a sufficient 
>> accuracy or can't you afford a larger basis set? For PBE, one should employ 
>> a basis set of at least double-zeta, maybe even triple-zeta quality.
>> - if you are already using smearing, you could use the actual temperature 
>> as the electronic temperature (seems more physical to me)
>> For more advanced suggestions, people more experienced in MD are open to 
>> jump in.
>> Best,
>> Frederick
>>
>> Michela Benazzi schrieb am Freitag, 26. Juli 2024 um 16:18:24 UTC+2:
>>
>>> Hello, Can anyone please give me their opinion/advice? I am trying to 
>>> run with TIMECON 50 (it was 1000 before), but no improvements yet.
>>>
>>> On Tuesday, July 23, 2024 at 7:23:56 AM UTC-4 Michela Benazzi wrote:
>>>
>>>> Hello everyone,
>>>>
>>>> I am running DFT at 1200 K, 1 bar with a periodic system of 64 Al atoms 
>>>> + 2 CO2 atoms. However, I am seeing my temperature slowly increase to >4000 
>>>> K and my pressure immediately spiking to ca. 2*10^5 bar.
>>>>
>>>> I was wondering if you could help me diagnose this issue - I thought 
>>>> perhaps it was because I was not accounting for vdW interactions, but I 
>>>> just added that term to my newer simulations and I am still incurring in 
>>>> the same problem!
>>>>
>>>> I included below my input and shortened output.
>>>>
>>>>

-- 
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 on the web visit https://groups.google.com/d/msgid/cp2k/b0964eac-152a-4837-9154-08a5ac66e9cbn%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240726/8c016704/attachment.htm>