Re: some recommendations for parametrizing new molecules required

From: Christopher Gillespie (gillescche_at_gmail.com)
Date: Fri Jan 23 2009 - 09:37:17 CST

Hi Axel,

My MP2 run is not for parameterization but something else entirely.
Yes I did pre-optimize at HF/6-31g*. I have played with a number of
basis sets based on the needs of the system although no I have not
done DFT as a comparison, so thanks for the idea.

Best
Chris

On Jan 23, 2009, at 10:25 AM, Axel Kohlmeyer wrote:

> On Fri, 23 Jan 2009, Christopher Gillespie wrote:
>
> CG> Hi Vlad,
>
> chris,
>
> CG> I have recently started a parameterization of my own and here is
> what I have
> CG> seen thus far.
> CG>
> CG> GAMESS supposedly has decent parallelization and potentially can
> be faster
> CG> than Gaussian. I say supposedly because I have only tried it
> once and have
> CG> used Gaussian for everything else. Gaussian has parallelization
> as well but
> CG> I have been informed GAMESS does a better job of it. Also, I
> have used the
> CG> US version and have not tried another.
> CG>
> CG> As for time, I have a 43 atom optimization (MP2/cc-PVDZ) with
> counterpoise
> CG> correction running for the last 13 days with Gaussian, which is
> higher than
> CG> the recommended HF/6-31g* level for CHARMM FFs but still your
> system is
> CG> quite a bit larger, QM speaking.
>
> now that looks like a nice waste of computer time. have you at
> least preoptimized the structure with HF/6-31g*? have you
> checked how much of a difference you would get by cranking
> up the basis set? have you considered, that in many cases
> hybrid DFT gives you similar "quality" results than MP2 at
> a fraction of the cost, particularly with larger systems,
> due to different scaling behavior. have you considered, that
> the extra level of theory is of limited use, since you want
> to fit a classical potential and the resulting error will
> probably outweigh the difference in error between the HF
> and the MP2 run?
>
> cheers,
> axel.
>
>
> CG>
> CG> Hopefully someone with more GAMESS experience can supply more
> information on
> CG> its speed and use with Paratool.
> CG>
> CG> Best
> CG> Chris
>
> --
> =
> ======================================================================
> Axel Kohlmeyer akohlmey_at_cmm.chem.upenn.edu http://
> www.cmm.upenn.edu
> Center for Molecular Modeling -- University of Pennsylvania
> Department of Chemistry, 231 S.34th Street, Philadelphia, PA
> 19104-6323
> tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425
> =
> ======================================================================
> If you make something idiot-proof, the universe creates a better
> idiot.

----------------------------------------------------------------------------------------
| Christopher Gillespie
| Ph.D. Candidate
| University of Delaware
| Department of Chemical Engineering
| 150 Academy Street
| Newark, DE 19716
| gillesc (at) udel.edu
| http://udel.edu/~gillesc
----------------------------------------------------------------------------------------

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:50:25 CST