AMBER Archive (2009)
Subject: RE: [AMBER] PMEMD 10 overflow issues
From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Sat Aug 22 2009 - 12:21:30 CDT
>> I should have said that I am getting corrupted restart files for the
>> explicit solvent simulations, due to what I think is overflow (******
>> values should be)
> Nothing is happening there. It is just that those parameters or values
> too high. So try to avoid such things, for you to understand what is
> going on.
> For electrostatic and non bonded interactions, try using scee and scnb.
Again with all due respect, this advice is incorrect. Firstly the poster is
referring to the coordinate values in the restart files and the fact that
this is occurring here but not in the trajectory file implies something
strange is happening.
With regards to your second comment, that of their being *'s in the EEL and
VDW terms and that "It is just that those parameters of values are too
high". Implies that such situations are not of great concern. This is
INCORRECT. If you EVER see *'s for energy terms in the output you have
serious issues with your simulation that will require addressing before any
meaningful data can be obtained.
Additionally with regards to the values of scee and scnb I will reiterate
the point that has been made many times before on this list. "DO NOT CHANGE
SCEE OR SCNB UNDER ANY CIRCUMSTANCES UNLESS YOU KNOW EXACTLY WHAT YOU ARE
DOING!" These values are force field specific and for all modern protein
force fields should be left at the defaults and NOT changed. Unless you are
using the extremely old Weiner et al force field, using the GLYCAM force
field, or developing your own specialized force field you should never have
reason to adjust these values.
All the best
| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |
Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.
AMBER mailing list