Parallel and Serial computation

Dear Axel

Thanks alot for the patient reply to each point. Sorry about the hydrogen half ions. I added that just to electrostatically neutralize the system. I read in wiki the following

Thus, the method (Ewald/PME) is best suited to systems that can be simulated as infinite in spatial extent. In molecular dynamics simulations this is normally accomplished by deliberately constructing a charge-neutral unit cell that can be infinitely “tiled” to form images; however, to properly account for the effects of this approximation, these images are reincorporated back into the original simulation cell.The electrostatic energy of a polar crystal (i.e., a crystal with a net dipole \mathbf{p}_{uc}in the unit cell) is conditionally convergent, i.e., depends on the order of the summation.

So to avoid this, I added an extra charge. I asked a friend of mine who suggested this. To be frank, I myself was not convinced by this idea, but thought of giving a try. I tried without this as well. The other system was not even equilibrating.

Thanks for your info regarding the chaotic effect.

I just wanted to see whether I am doing a parallel computation correctly. Even though it is just a matter of submitting a script into the cluster (LAMMPS made life easy :D) , I wanted to be sure of the fact that I am doing things not weirdly. The example I started with is definitely looking weird the way I proceeded. So I thought of looking at a standard example first and hence PRD. I am really sorry to bother you quite alot. Thanks again.

Regards

Vaidyanathan M S

From: Vaidyanathan M.S <[email protected]>
Date: Tue, Jul 16, 2013 at 10:01 AM
Subject: Re: [lammps-users] Parallel and Serial computation
To: Axel Kohlmeyer <[email protected]>

Dear Axel

Thanks alot for the patient reply to each point. Sorry about the hydrogen
half ions. I added that just to electrostatically neutralize the system. I
read in wiki the following

in _what_ wiki?

Thus, the method (Ewald/PME) is best suited to systems that can be
simulated as infinite in spatial extent. In molecular dynamics<http://en.wikipedia.org/wiki/Molecular_dynamics>simulations this is normally accomplished by deliberately constructing a
charge-neutral unit cell that can be infinitely "tiled" to form images;
however, to properly account for the effects of this approximation, these
images are reincorporated back into the original simulation cell.The
electrostatic energy of a polar crystal (i.e., a crystal with a net dipole [image:
\mathbf{p}_{uc}] in the unit cell) is conditionally convergent<http://en.wikipedia.org/wiki/Conditionally_convergent>,
i.e., depends on the order of the summation.

this obviously applies to slab systems. if you look through publications
studying diffusion of a single ion in liquids, you'll predominantly see
simulations of charged systems. while it is in general a good idea to have
a neutral system, your workaround is worse than the problem. you add
completely unphysical particles for which you have no parameterization to
your system that float around and can distort the structure or cluster with
your ion and whatnot. a charged system will be automatically neutralized
with a compensating background charge (simply by ignoring the divergent
part at k=0).

So to avoid this, I added an extra charge. I asked a friend of mine who

suggested this. To be frank, I myself was not convinced by this idea, but
thought of giving a try. I tried without this as well. The other system was
not even equilibrating.

i don't understand what you are saying here.

Thanks for your info regarding the chaotic effect.

I just wanted to see whether I am doing a parallel computation correctly.
Even though it is just a matter of submitting a script into the cluster
(LAMMPS made life easy :D) , I wanted to be sure of the fact that I am
doing things not weirdly. The example I started with is definitely looking
weird the way I proceeded. So I thought of looking at a standard example
first and hence PRD. I am really sorry to bother you quite alot. Thanks
again.

the closest example to compare to are either the "peptide" example or the
"rhodo" benchmark.

axel.