[gmx-users] NVT-Eq ---> NVT-MD

[Justin Lemkul]

On 2/15/16 12:40 PM, Alexander Alexander wrote:
> Dear Mark,
>
> To be honest, I do not know exactly when and in which physics the velocity
> generation in a MD simulation is necessary and when not !
>
> And I usually-historically generate velocity in the Equilibration and not
> in Production.
>
> My specific case here is a pure metallic surface.
>

Here is a simple decision structure:

Do the atoms have velocities?
If yes, don't re-generate them.
If no, you need to generate them.

gen_vel randomly assigns velocities according to a Maxwell distribution.  If 
you're continuing a previous run, you already have velocities that are 
(presumably) part of the target statistical mechanical ensemble.  Re-generating 
them destroys that and requires re-equilibration.  If you're just starting out, 
you've probably done an energy minimization, which has no velocities.  So if you 
try to start a simulation without generating velocities, your system may behave 
rather strangely as the thermostat tries to figure out how to bring the 
non-existent velocities into agreement with the target temperature.

-Justin

> Regards,
> Alex
>
> On Mon, Feb 15, 2016 at 6:27 PM, Mark Abraham <mark.j.abraham at gmail.com>
> wrote:
>
>> Hi,
>>
>> Why did your two-part run need to generate velocities at a suitable
>> temperature in the first part? Does that reason apply to your one-part run?
>> ;-)
>>
>> Mark
>>
>> On Mon, Feb 15, 2016 at 6:23 PM Alexander Alexander <
>> alexanderwien2k at gmail.com> wrote:
>>
>>> Thanks for your prompt response,
>>>
>>> Could you please confirm me that  in my single NVT-MD I should  use
>>> "continuation  = no" and below velocity generation, right? or velocity
>>> generation is not required any more?
>>>
>>> ;Generate velocity
>>> gen_vel = yes
>>> gen_temp = 298.15
>>> gen_seed = -1
>>>
>>> Regards,
>>> Alex
>>>
>>> On Mon, Feb 15, 2016 at 6:12 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>>>
>>>>
>>>>
>>>> On 2/15/16 12:09 PM, Alexander Alexander wrote:
>>>>
>>>>> Dear Gromacs user,
>>>>>
>>>>> I have an elementary question which I would be thankful if one could
>>>>> comment on it;
>>>>>
>>>>> I want to do a NVT-MD simulation as a production calculation of my MD
>>>>> simulation for 10 ns, and suppose that a NVT- Equilibration has been
>>> done
>>>>> in advance for 0.5 ns. The normal differences between their .mdp files
>>> are
>>>>> coming below and the rest of their .mdp files are fully identical;
>>>>>
>>>>> NVT-Eq (Equilibration)
>>>>> ------------------------------------
>>>>> nsteps  =  500000    ; 0.5 ns
>>>>>
>>>>> ;Generate velocity
>>>>> gen_vel = yes
>>>>> gen_temp = 298.15
>>>>> gen_seed = -1
>>>>> -------------------------------------
>>>>>
>>>>>
>>>>> NVT-MD (Production)
>>>>> -------------------------------------
>>>>> nsteps = 10000000    ; 10 ns
>>>>> continuation  = yes    ;continuing from NVT-Eq
>>>>>
>>>>>
>>>>> ;Generate velocity
>>>>> gen_vel = no
>>>>> -------------------------------------
>>>>>
>>>>> Since they have identical .mdp parameters, then, my question is that:
>>>>>
>>>>> if it is possible to do just one NVT-MD simulation (a bit longer (10 +
>>>>> 0.5)ns ) instead of two sequence NVT-Eq and NVT-MD simulations?
>>>>>
>>>>>
>>>> Of course.  The normal use for separating two phases (equilibration and
>>>> production) is in the case of restraints, shorter time step for
>>>> equilibration in troublesome systems, etc.  If you just have the same
>>>> settings, do a single run.
>>>>
>>>> -Justin
>>>>
>>>> --
>>>> ==================================================
>>>>
>>>> Justin A. Lemkul, Ph.D.
>>>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>>>
>>>> Department of Pharmaceutical Sciences
>>>> School of Pharmacy
>>>> Health Sciences Facility II, Room 629
>>>> University of Maryland, Baltimore
>>>> 20 Penn St.
>>>> Baltimore, MD 21201
>>>>
>>>> jalemkul at outerbanks.umaryland.edu | (410) 706-7441
>>>> http://mackerell.umaryland.edu/~jalemkul
>>>>
>>>> ==================================================
>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>> posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>> send a mail to gmx-users-request at gromacs.org.
>>>>
>>> --
>>> Gromacs Users mailing list
>>>
>>> * Please search the archive at
>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>> posting!
>>>
>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>
>>> * For (un)subscribe requests visit
>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>> send a mail to gmx-users-request at gromacs.org.
>>>
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> send a mail to gmx-users-request at gromacs.org.
>>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================