Hi Gustavo:
All that you told is very interesting and useful. Thanks. However, I would like to see interfaces to NWChem too. Think that NWChem is completely free for non-commercial use and I believe - as a NWChem user - it can do the job that Gaussian does, if you rule out some illusions from the latter, as getting the IR of the compounds like from an IR spectrometer, or calculating the CD of compounds in condensed phase. Also, the binaries that NWChem offers, while quite fast, could even be much faster if compiled with "faster" compilers (NWChem is in Fortran and as offered in binaries is compiled with f77 or gfortran, which are much the same) with vendor math libraries (may be acml, which are also free). I plan to carry out such compilation soon. Many can not have the money for Gaussian. And the situation may subtract energies to such codes as NWChem and MPQC, the latter perhaps the best of all, although underdeveloped.
But, these are naive ideas from an experimentalist who has approached the virtual only recently.
Regards
francesco
--- On Thu, 5/22/08, Gustavo Seabra <gustavo.seabra.gmail.com> wrote:
> From: Gustavo Seabra <gustavo.seabra.gmail.com>
> Subject: Re: AMBER: $AMBERHOME/dat/slko
> To: amber.scripps.edu
> Date: Thursday, May 22, 2008, 10:00 AM
> Hi Francesco,
>
> > I had downloaded the new files of data. Do you know if
> improved?
>
> As Ross said, no. The parameters are exactly the same. The
> changes in
> the files are just the naming conventions, and the
> additional xml
> section in the end with a description of how parameters
> were obtained.
> (This xml session is just ignored in the calculations.)
>
> > With the old parameters and Amber9 I could do little
> for highly conjugated
> > compounds: the dftb minimized conformations were far
> from reality, much
> > like with PM3 and the other semiempiricals. Actually,
> even B3LYP proved
> > inadequate, [...]
>
> That alone provides a hint that DFTB would probably not
> work well. If
> you look in the xml part of the new files, you'll
> notice that all of
> them used B3LYP/6-31G* in the calculation of the repulsive
> potentials.
>
> > so that I had to stick to vacuum (as COSMO and the
> others
> > also proved inadequate for the tiny differences in
> conformational energies):
> > F. Pietra "Why colchicine does not show
> mutarotation. With M05‑2X density
> > functional in the realm of tricky natural products. J.
> Phys. Org. Chem. 2007,
> > 20, 1102‑1107"
>
> So, did M05-2X work fine? If so, you it may be interesting
> for you to
> take a look at the new Amber/PUPIL/Gaussian interface
> available in
> Amber10. Take a look at:
>
>
> http://www3.interscience.wiley.com/journal/117910767/abstract
>
> Where this interface is discussed.
>
> > At any event, if I understand, the only file for
> typical use in Amber10
> > is is the SCC file mio.0-1-tar.gz (S-O-N-C-H), the one
> I have used for slko.
>
> Yes, that should cover most applications. However, Amber10
> does not
> limit the possible atoms anymore, so you can use whatever
> atoms you
> want, as long as you have the needed slko files. That
> allows you to
> download any parameter set form dftb.org and use it with
> Amber10. Just
> make use that, if you need to use files from different
> sets, the sets
> are compatible.
>
> > I take also the opportunity to ask if Amber9 can be
> removed or if it
> > might be useful alongside Amber10 (neglecting
> comparisons, of course).
>
> There have been some changes in output format from Amber9
> to Amber10.
> Right now REMD comes to mind. So, you *may* want to keep
> Amber9 around
> just until the users get comfortable with the new formats.
> (But that's
> just *my personal* opinion.)
>
> Gustavo.
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the
> *body* of the email)
> to majordomo.scripps.edu
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
to majordomo.scripps.edu
Received on Sun May 25 2008 - 06:07:34 PDT
