<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#ffffff">
On 8/12/2010 4:52 AM, Igor Marques wrote:
<blockquote
cite="mid:AANLkTi=XbrXJa0L-1BCVwjDgii9TK09eLj7i9KY-39K6@mail.gmail.com"
type="cite">Hello everybody,<br>
<br>
I've digging around the user manual, the website and the mailing
list archives and I'm afraid I might be missing something: does
GROMACS have some kind of <i>neutral plasma </i>as AMBER does?<br>
</blockquote>
<br>
You're probably referring to some descriptions of the technique
underlying the Ewald method.<br>
<br>
<blockquote
cite="mid:AANLkTi=XbrXJa0L-1BCVwjDgii9TK09eLj7i9KY-39K6@mail.gmail.com"
type="cite">
Or every simulation must have a net charge of 0.000 ?<br>
</blockquote>
<br>
All the Ewald derivations I have seen require neutrality of the
simulation cell, however that requirement can finessed by
introducing a neutralizing plasma for each charged entity. I learned
about this issue a while ago in a namd-l thread
<a class="moz-txt-link-freetext" href="http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/3976.html">http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/3976.html</a>.
The reference Hyonseok cites (T. Darden, D. Pearlman, L. G.
Pedersen, JCP v109, p10921, 1998) covers the issue well, and the
salient point for MD in GROMACS is that the self potential in (2.16)
vanishes under the differentiation that forms the forces from the
energy. So the forces are right regardless. That reference does
post-date the SPME paper that GROMACS implements. I'm not sure if
GROMACS has the correct self-energy for non-neutral cells, but that
only affects the total energy as an additive constant.<br>
<br>
Perhaps Berk or Carsten can enlighten us about the self-energy
implementation.<br>
<br>
Mark<br>
</body>
</html>