[CP2K-user] different version of cp2k different energy

[Fabian Ducry]

Hi Xin,

It depends on the system you are simulating, e.g. molecules and 2D 
materials often profit from a vdw correction. For bulk materials however, 
the correction may lead to too short bond lengths. Make sure to check for 
your input. Also compare different vdw corrections.

Fabian

On Tuesday, 28 July 2020 21:31:24 UTC+2, Xin Wei wrote:
>
> Hi Fabian, 
>
> Thank you very much! I didn't notice that one input file doesn't have VDW 
> correction before. I tried the same input file for different versions of 
> CP2K and the results are close. It's better to include the VDW correction 
> in the calculations, right? 
>
> Xin
>
> On Tuesday, July 28, 2020 at 3:03:01 AM UTC-4, Fabian Ducry wrote:
>>
>> Hi Xin,
>>
>> Have you tried to run both input files with the same version of cp2k? 
>> That would be the easiest test whether there is a problem or not. 
>> The two input files are quite different in fact: You are using two 
>> exchange-correlation functionals: one has a VDW correction while the other 
>> does not. So you cannot expect to get the same energy, not even similar 
>> values.
>>
>> Best,
>> Fabian
>>
>> On Tuesday, 28 July 2020 01:22:29 UTC+2, Xin Wei wrote:
>>>
>>> Hi Everyone, 
>>>
>>> I did the geometry optimization using similar input files for the same 
>>> structure but different versions of cp2k --cp2k6.1 and cp2k7.1. (The input 
>>> files are attached) The final energy calculated is different and the 
>>> difference is -339.071 KJ/mol. I am wondering what caused this difference. 
>>> Are different versions of cp2k using different reference energy? If so, may 
>>> I know what is the reference energy? 
>>>
>>> Thanks a lot!
>>>
>>> Xin
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200728/16ffcf33/attachment.htm>