I think we talk about two different issues.
There is absolutely no objection for cases where things have to be added
to
accommodate extra features either within sander or anywhere else. Of
course,
we hope that old things are not broken by doing so or if things have to be
broken, document it. Both PIMD and NEB fall within this category. They
both
have been developed for important new features and they do not affect any
old things.
What I want to know is to what extent the "re-building" excluded list
within
sander affects other part of the code. For example, if I add a pair of
atoms
on the excluded list in prmtop that are not connected by three or fewer
bonds, does sander take it. If not, what should I do? Can I avoid playing
with the code?
Tom, can you give us a bit explanation on this, please. It seems we all
have
fired back and forth a lot of rounds by now, but you have been so quiet.
Can
you tell us what has been done, specifically? Does it affect anything
other
than extra points?
This may have implications outside sander (e.g., energy-analysis modules).
Thanks!
yong
>
Note that the PIMD and NEB options also modify the excluded atom list that
is
present in the prmtop file. Given that the excluded atom "list" can
become
too large to be stored in the prmtop file, or even to be in memory, such
accommodations seem inevitable.
....dac
Received on Wed Apr 05 2006 - 23:49:49 PDT
