On Fri, Sep 05, 2008, Robert Duke wrote:
> I am fine with what Ross proposes here; it may make sense to study the best
> value for dsum_tol a bit more, but I have a hard time seeing how we could
> lose if we just decrease it from the current default by a factor in the
> range of 3.3... to 10 (ie., we would go no lower than 1.e-6).
I'd be in favor of changing the default in Amber11 to 1.e-6.
> I would be
> curious if Darden has any words of wisdom on the original selection of
> 1.e-5, in case we are missing something here.
I've tried to get Tom to commit on this, but he keeps saying things like
"oh, it's complicated", but I don't think has had time to really study
the problem anew.
...dac
Received on Fri Sep 12 2008 - 03:07:24 PDT
