On Mon, Apr 03, 2017, Chris Neale wrote:
> I was unable to understand the actual intended meaning in David's post, but
> I still think an error message would be good. If iwrap=1 is possible and
> leads to unexplained crashes with restraints, then a sensible error message
> would help the user.
In most cases, positional restraints are applied to the biomolecule part of
the system, which is often not wrapped. So it is often OK to set iwrap=1
when ntr>0. I'm not sure what caused the bad interactions on restart in this
particular case (maybe it was in some email I didn't read carefully enough.)
> at some point in the
> future unwrapped coordinates are going to exceed floating point capacity.
I'm not holding my breath here: the range of even single=precision floats
is pretty impressive. The iwrap option was put in before we supported netcdf
restart files. I think we should discourage its use now. If you are worried
about these issues, then autoimage your restart file after every microsecond
or so (or more often if you wish).
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Apr 03 2017 - 19:00:03 PDT
