[CP2K-user] [CP2K:18063] starting a AI MD simulation from minimised structure

[jing liang]

there are about 900 atoms in the simulation box

El lunes, 21 de noviembre de 2022 a las 11:03:54 UTC+1, nihar... at gmail.com 
escribió:

> First of all, how big (no. of atoms) is your system?
>
> If it is fairly small (around 500 -700 atoms with some 80 processors) then 
> you can try to attempt AIMD simulation.
>
> Best thing, I found, to start an AIMD simulation is to run a few pico 
> seconds (may be 100-200 ps) classical simulation with a reasonable force 
> field.
> Then start your AIMD from the last step configuration.
>
> Best of luck.
>
> N Choudhury
>
> On Monday, November 21, 2022 at 2:17:54 PM UTC+5:30 jingli... at gmail.com 
> wrote:
>
>> Thank you for your useful comments, I am starting to think about 
>> returning to namd/gromacs. But before
>> I do it, I wanted to ask you what criteria do you use to find out whether 
>> the system is equilibrated or not. 
>> I found this page https://www.cp2k.org/howto:md?do=  From it, I 
>> understood that by looking at the column
>> "Cons Qty[a.u.]" from the .ener file, one could get the difference 
>> between the maximum and minimum of the
>> curve and divide it by the kinetic energy, for an equilibrated system 
>> this ratio should be < 1% Maybe I didn't
>> get the point? 
>>
>>
>> El lunes, 21 de noviembre de 2022 a las 0:38:09 UTC+1, 
>> dresear... at gmail.com escribió:
>>
>>> Ideally you need to use force field to optimize your structure 
>>> classically. I do this for liquids.
>>> Then, you can continue NVE, T by quick step CP2K. Equilibration with 
>>> AibMD is very time consuming, you will unlikely be able to equilibrate 
>>> random structure with >300-500 atoms, or you could but it will be very 
>>> expensive and long simulation. I couldn't equilibrate NPT of random liquid 
>>> of 16*31+32*18+32 atoms with 36-48 cores for more than a month.
>>>
>>>  
>>>
>>>
>>>
>>> On Mon, Nov 21, 2022, 1:18 AM jing liang <jingli... at gmail.com> wrote:
>>>
>>>> I am trying to see how the structure evolves in time, and monitor where 
>>>> some bond rotation occurs. The long story short, I
>>>> could pack all relevant molecules in a box. I was doing some search for 
>>>> parameters for gromacs/namd/... but I found it
>>>> not straightforward, also the mixing of force fields is something I am 
>>>> not completely sure about. Then, I thought that pure
>>>> CP2K simulation could work, and it is doing a great deal in the 
>>>> minimisation process. But I don't know how long it could
>>>> take to converge completely. 
>>>>
>>>> As I wrote before, the goal of this simulation would be to see if some 
>>>> bond rotation occurs in the long term some spectra
>>>> would be great but at this point just the dynamics. That's why I asked 
>>>> here if there is some standard/accepted criteria
>>>> for using a structure as the initial frame for an AI MD simulation 
>>>> regarding the energies/forces. Thanks for you comments. 
>>>>
>>>>
>>>>
>>>> El domingo, 20 de noviembre de 2022 a las 15:40:11 UTC+1, 
>>>> dresear... at gmail.com escribió:
>>>>
>>>>> This is a good question but it depends on what you are simulating. If 
>>>>> it is non- equilibrium system then it is pointless to look at the energy, 
>>>>> temperature, pressure etc.
>>>>>
>>>>> If it is equilibrium system, it should be similar to classical MD but 
>>>>> on a shorter time scale. There is a tutorial how to calculate IR spectra 
>>>>> with cp2k, you can find steps there for equilibration and spectrum data 
>>>>> acquisition. 
>>>>> It is on cp2k website.
>>>>>
>>>>>
>>>>> On Sun, Nov 20, 2022, 5:33 PM jing liang <jingli... at gmail.com> wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I created a simulation box manually and then started a minimisation 
>>>>>> process
>>>>>> by using geometry optimisation. My goal is not to get a fully 
>>>>>> optimised structure
>>>>>> but to start a MD simulation using the NVE ensemble. My question is 
>>>>>> if one needs
>>>>>> a fully optimised structure in order to start MD or if there is a 
>>>>>> criterion to say
>>>>>> that the minimised structure is already fine? 
>>>>>>
>>>>>> I am coming from classic MD simulations where some software suggest 
>>>>>> to have
>>>>>> the forces converged up to some particular threshold. 
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> -- 
>>>>>> 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+uns... at googlegroups.com.
>>>>>> To view this discussion on the web visit 
>>>>>> https://groups.google.com/d/msgid/cp2k/7c1fb302-ae85-4a70-88d6-8dca9be6849en%40googlegroups.com 
>>>>>> <https://groups.google.com/d/msgid/cp2k/7c1fb302-ae85-4a70-88d6-8dca9be6849en%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>>> .
>>>>>>
>>>>> -- 
>>>> 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+uns... at googlegroups.com.
>>>>
>>> To view this discussion on the web visit 
>>>> https://groups.google.com/d/msgid/cp2k/8c294cd6-a99a-482c-8a84-304cbbbee083n%40googlegroups.com 
>>>> <https://groups.google.com/d/msgid/cp2k/8c294cd6-a99a-482c-8a84-304cbbbee083n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>> .
>>>>
>>>

-- 
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/36335f26-5514-4e60-b368-526249f890f3n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20221121/9aa28e21/attachment.htm>