On Sun, May 04, 2014, 张明焜 wrote:
>
> normal mode analysis
> &data
> ntrun=1,
> maxcyc=0,
> nsave=250, nprint=250,
> cut=10.0,
This is dangerous, since the cutoff is very short for a vacuum simulation.
Almost always, you should set this value to some very number, such that
no interactions are ignored.
>
> scnb=2.0, scee=1.2,
Be sure these values are the same as in sander.
> Then check Each energy component of sander result and nmode result,I found
> the E-EEL14 item is different, which is 38.2626 in sander. What's this
> about? So this is my first queston.
Are all other parts of the energies the same? (especially the total
electrostatic energy?) Make sure the dielectric behavior is the same in
the two programs -- I think the defaults may be different.
[Note that nmode is really obsolete: consider using the nmode() function in NAB
instead.]
>
> Later ,I turn up the drms to 5.0 to to avoid this error,
No...you can't do this. You need to have a very low RMS value in nmode,
at least below 10**-5 kcal/mol-A.
I don't know what happened to your eigenvectors. Make sure the test cases in
$AMBERHOME/test/nmode pass; see if you can see differences between those
inputs (which do produce eigenvectors) and you example is.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun May 04 2014 - 08:30:03 PDT
