"Bizarre results" is surprising.... unless... your first trajectory is in
netcdf or binary format. IMIN=5 does not yet support netcdf. If this
sounds like it could be a problem, you might try reading the trajectory in
with ptraj, and writing it out, unmodified, with box information, in
ASCII.
A few years ago, I used IMIN=5 for free-energy perturbation work (which
required prcessing solvated octahedron box trajectories with two (or
three) slightly different prmtop files). IMIN=5 worked well for this -
and these trajectories had many waters leaving the periodic cell. So, I
am a little surprised by these troubles.
IMIN=5 certainly needs to work properly - for a wide variety of important
applications.
Finally, a precise analysis of _what_ energies are different between
initial sander and post-process with imin=5. Knowing that all the bonded
terms are identical (or nearly so due to my other caveat) would be
encouraging. If the ENERGY issues could be attributed to just one or two
comopnents of total energy that might be helpful.
Chris
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Dec 02 2007 - 06:07:39 PST
