HI,
I agree that the relative energies are the real relevant information,
and this is comforting from the developer's point of view because at
least previous work is not invalidated.
However, there seems to be no disadvantage on moving to more precise
values for the physical constants and conversion factors, aside from
the fact that results may not match previous work, which may be a
problem for the test cases. In this case, some "use_old_constants"
option could solve this last problem.
Cheers,
Gustavo Seabra
Professor Adjunto
Departamento de Química Fundamental
Universidade Federal de Pernambuco
Fone: +55-81-2126-7417
On Fri, Sep 3, 2010 at 10:34 AM, Ben Roberts <roberts.qtp.ufl.edu> wrote:
> Hi Adrian,
>
> Yes. I agree that absolute energies are not necessarily informative, and that it takes a rather large absolute energy for the error to manifest in terms of kcal/mol.
>
> I guess, then, the only other question is that of forces - but again, I suppose, the QM force would have to be very large, both absolutely and relative to the other components of the force, for the difference to become important. Now that I look at it, I suppose two parts in 100,000 is not such a big deal.
>
> Ben
>
> On 2/9/2010, at 1:50 p.m., Adrian Roitberg wrote:
>
>> Dear Ben
>> I would make the strong claim that the absolute energy is pretty
>> irrelevant. I would only look at relative energies. In that case, while
>> your argument still holds in principle, it is clear that we will be
>> talking about meV and not millions...
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 03 2010 - 07:00:06 PDT
