AMBER Archive (2007)Subject: Re: AMBER: antechamber, how does it work
From: Seth Hayik (hayik_at_ufl.edu)
Date: Wed Aug 15 2007 - 10:10:22 CDT
I'm not an expert on antechamber either, but I'll try to shed
whatever additional light I can.
Firstly, most everything that Gustavo said is correct, AM1 is just
a Hamiltonian that Divcon uses. From what I gather what you're
trying to do with antechamber is completely separate from the
QM/MM implementations you're trying, so SANDER won't get used at
all.
(1) Can divcon be launched selectively with amber (and an
important option for
>> RESTARTING should a crash occur), or AM1 is what can be
>> launched? It seems to
>> me that it is not the same thing.
Divcon is, at its heart, a stand-alone application. The version
included in SANDER has been stripped down to use only certain
keywords and methods that are needed in antechamber, so I don't
think that the included divcon would work properly.
(2) When launching antechamber as in tutorial B4 (the - more or
less - way I am
>> using it), is "divide and conquer" at work? I see, during
>> present antechamber
>> job, listed in divcon.out statements like "Resetting search
>> direction". What
>> does that mean? That guess Hessian is recomputed from the last
>> good
>> coordinates? Or does that mean that minimization of my molecule
>> is done by
>> formally cutting it into pieces (divide and conquer)? The way I
>> was used at
>> when computing charges "ab initio" via HF-6-31G* by formally
>> cutting the
>> molecule into residues.
As Gustavo said, "resetting search direction" has to do with the
minimizer, nothing with the divide and conquer approach. While
Divcon is capable of divide and conquer calculations I don't think
that these are default within antechamber. Divcon can do
standard, ie. non divide and conquer which is default, or divide
and conquer through keywords in the input file. As I said, I'm
not an antechamber expert so I'm not sure if there is a way to
turn on this functionality in antechamber. Also, the divide and
conquer algorithm doesn't exactly cut the molecule into pieces,
you can see more on how it works in: Dixon, S.L. and K.M. Merz, F
Journal of Chemical Physics, 1997. 107(3): p. 879-893. if you're
interested.
3) Amber9 manual, after stating that divcon is called by
antechamber (p 74),
>> refers to AM1 in the following pages, none of the many options
>> of divcon
>> (chapter 7) being mentioned.
As Gustavo said, this refers to using Divcon to do a calculation
with the AM1 Hamiltonian.
(4) When launching a QMMM procedure from sander.MPI, is the QM
part running
>> parallel? If so, why not applying parallelization for the AM1
>> minimization
>> within antechamber?
When you are using QM/MM you are probably using the native QM code
for your simulations and this is not related to using antechamber.
Divcon has some parallelization, but I don't know how it applies
in antechamber, and you shouldn't need it for a molecule this size
if the divide and conquer algorithm works within antechamber. I
don't know how/if that can happen, maybe someone that knows more
about antechamber can help you there. Another option might be to
loosen the minimization criteria by setting XTEST, ETEST and/or
GTEST to higher values at the expense of a less accurate minimized
structure. This can be done using the -mk keyword(p. 75 in the
manual), but you must specify the entire input file then, so you
might want to look at the other keywords already in the divcon.in
file.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
|