[gmx-users] Assignment of new velocities for multi-runs

[Justin Lemkul]

On 9/29/12 3:08 AM, Peter C. Lai wrote:
> If the system is already well equilibrated previously then you can probably
> safely reassign velocities during NPT. The best way to do this would probably
> be to take the equilibrated checkpoint file as the starting point:
>
> grompp -f genvel-yes-npt.mdp -t previous.cpt ...
>
> You'll still have to wait some ns for the system to reequilibrate so data
> comparisons between the runs for the first several ns will not be meaningful,
> so ultimately, the question is what you do you hope to gain from reassigning
> velocities at the beginning?
>

I disagree with the approach of reassigning velocities at the outset of NPT. 
Doing so can lead to instabilities in the pressure coupling algorithm and the 
system may crash.  The simplest answer is to just use different starting 
velocities at the very beginning of the workflow (i.e. velocity generation at 
NVT with different values of gen_seed).

-Justin

> On 2012-09-28 11:59:02PM -0700, James Starlight wrote:
>> Dear All!
>>
>> I'd like to perform several simulations of the membrane protein
>> started from the common conditions which differs only in the initial
>> velocities ( for each simulation random speed distribution will be
>> used from the Maxwell distribution).
>>
>> Because I simulate membrane protein the long equilibration phase is
>> required for that system. I've noticed that new velocities sets are
>> assigned in the NVT phase of equilibration (100ps) after which more
>> prolongest npt equilibration (10ns) is followed.
>>
>> Does it require that new velocities always be assigned exactly during
>> equilibration of my system ? Could I use new velocities in the
>> production md run instead ? In the latter case my system will be
>> equilibrated with one velocities but further production MD will be
>> with another velocities each time? (so in latter case I can avoid long
>> equilibration phase for each simulation).
>>
>>
>> James
>> --
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> * Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>

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

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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