# [CP2K-user] [CP2K:18197] Re: Equation used for annealing

Moon Moon yuegrit at gmail.com
Wed Dec 14 10:58:24 UTC 2022

```Hi, Lenard.
Have you resolved the problem of changing temperature in cp2k?
I also can't get expected results by using TEMPERATURE_ANNEALING.

> I might be wrong, but for the NVE ensemble, T(n)=s^n * T(n=0) can be
> derived directly from |v|(n)=q^n * |v|(n=0) if we write:
> T(n) = (Σ[k=0→K] |v_k(n)|^2)/(3k(K-m)) = (Σ[k=0→K] |q^n
> v_k(0)|^2)/(3k(K-m)) = q^n T(0)
>
> So I should then be able to use T(n)=s^n * T(n=0) to find the value of s.
> But when I tried this earlier, I got much different results. That's why I
> just wanted to make sure my understanding of n in T(n) is correct.
>
> In another case I am using a 20 fs time step and 1000 steps, equating to
> 20 ps. Is n still an element of [0,1000]?
>
> On Thu, Apr 8, 2021 at 2:37 PM Lenard Carroll <lenardc... at gmail.com>
> wrote:
>
>> And this n, is this per MD step? For example, if I run the simulation for
>> 20 ps with a time step of 2 fs [10000 steps], is n then an element of
>> [0,10000]?
>>
>> On Thu, Apr 8, 2021 at 1:09 PM Julian Heske <jjh... at gmail.com> wrote:
>>
>>> Dear Lenard,
>>>
>>> your equation is true for the keyword TEMPERATURE_ANNEALING, which
>>> modifies the target temperature of the thermostat each step and I would
>>> suggest you to use this. The more general form is than
>>>
>>> T(n)=s^n * T(n=0).
>>>
>>> However, the ANNEALING parameter is for the velocities, so my guess is
>>> (I never used it, please s.o. correct me if I'm wrong):
>>>
>>> |v|(n)= q * |v|(n-1)
>>>
>>> and
>>>
>>> |v|(n)=q^n * |v|(n=0).
>>>
>>> Using the relation T(n) ~ |v|(n)^2 (
>>> https://en.wikipedia.org/wiki/Thermal_velocity ), you can also
>>> calculate the impact of the scaling parameter q on the temperature if you
>>> want to use this one.
>>>
>>>
>>>
>>> Best,
>>> Julian
>>> ASSIDUO Network schrieb am Dienstag, 6. April 2021 um 08:43:25 UTC+2:
>>>
>>>> Hi there everyone. Does anyone perhaps know the equation that is used
>>>> for annealing (https://manual.cp2k.org/trunk/CP2K_INPUT/MOTION/MD.html).
>>>> Here I know a scaling factor is multiplied to the initial temperature, but
>>>> I don't know if it takes on the form of:
>>>> T(n) = sT(n-1) or some other complicated equation.
>>>>
>>>> Could anyone let me know as I want to do the annealing properly.
>>>>
>>>> Much appreciated,
>>>> Lenard
>>>
>>> --
>>> 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
>>> .
>>>
>>

--
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.