On Thu, Apr 4, 2013 at 4:36 AM, <nguyentb.bii.a-star.edu.sg> wrote:
> Dear all AMBER expert,
>
> I'm an new user in AMBER, and hence, I have a lot of downs when
> choosing parameter for particular ensemble.
> I want to run NVE simulation, the paramaters I set are following:
>
As written, your mdin file will not be recognized by sander or pmemd.
Comments are indicated using the Fortran-style comment (i.e., replace #
with !)
> prod
> &cntrl
> imin=0, # turn off minimixation = md
> irest=1, ntx=5, # restart calculation, requires velocities in
> coordinate input file
> nstlim=5000000, # number of MD steps
> dt=0.002, # time step (ps)
> ntc=2, # flad for SHAKE, bonds involvimg hydrogen are contrained
> ntf=2, # force evaluation, bond interactions involing H-atoms omitted
> cut=14.0, # specific the nonbonded cutoff, in Angstroms
> ntb=1, # periodic boundary, constant volume
> taup = 0.1, # pressure relaxation time (in ps)
> ntpr=5000, # every ntpr steps energy information will be printed in
> human-readable form to mdout file
> ntwx=5000, # every ntwx steps the coordinates will be written to mdcrd
> file
> ntt=0, # pure Newtonian dynamics
> ig = -1 # change the random number seed from the default on every MD
> restart
> temp0=300.0, # reference temperature
> /
>
>
> Would you mind telling me whether the parameter is suitable for NVE
> ensemble?
It seems to be reasonable. You should check the quality of the energy
conservation (you may want to print out energies more often than every 5000
steps to do so). If you see significant energy drift, you may want to
reduce the time step to 0.001 (1 fs instead of 2).
> If I choose taup = 0.1, does my simulation change to NPE or
> NPV?
>
Neither. To turn on constant pressure, you need to use the "ntp" variable
(1 for isotropic pressure scaling and 2 for anisotropic pressure scaling).
I've never heard of anybody running NPE before (since compressing the unit
cell will not conserve energy). If you turn on a Barostat, you really
should turn on a thermostat as well. NPV, on the other hand, is not an
ensemble at all. Since P and V are conjugate variables, you cannot hold
both constant at the same time. Same goes for N and mu (chemical
potential) and T and E.
Do I need to set the ntp=0 and tempi=300.0? How about iwrap = 1,
> tol=0.000001 (SHAKE) and dsum_tol to 10**-6 ?
>
ntp=0 is the default (and means constant volume). So setting ntp=0 and
omitting it from the input file altogether results in exactly the same
simulation. iwrap=1 has to do with what images of atoms are printed in the
coordinate file. If iwrap=1, then the images inside a single unit cell are
always tracked (i.e., if an atom exits the unit cell, then the coordinates
being printed are then switched to its periodic image that just entered the
unit cell). Allen and Tildesley give a good description of periodic
boundary conditions if this concept is still confusing.
As for the tolerances, I have never had a problem just using the default
values.
Good luck,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Apr 04 2013 - 05:00:03 PDT
