On Thu, Jul 14, 2011, Qian Wang wrote:
>
> I am using amber10 to run MD. I built both old prmtop (using amber 6)and
> prmtop 7 for my system. For energy minimization, both old prmtop and
> prmtop give you very similar result. However, if I run MD for 100,000
> steps (using the same random number seed), the distribution of potential
> energy shifts several kcal/mol. Is there any reason for this? Thanks
> very much.
Please see section 1.3.4 of the Amber Users' manual. You should probably
run 20-50 steps of MD with each, setting ntpr=1, and see if you get identical
results. If you don't, you need to try to track down where the differences
are coming from. [It's not clear what parameters you are using, why the
"random seed" would make any difference, or what you mean by "very similar
result" in the email above. For the same coordinates, you should get the same
energies (to the level of precision printed in the output files), or should
be able explain where the differences are coming from.]
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 14 2011 - 13:30:03 PDT
