Re: PME fails no matter what

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Tue Jul 29 2008 - 03:18:21 CDT

On Tue, 29 Jul 2008, Roman Petrenko wrote:

RP> serched archive, changed addition to basis vectors to 0.25A -- still,
RP> no solution to the problem. it runs first 500 steps and gives the
RP> error.

so run for 400 steps. check what happens by writing out the trajectory
in every step. probably some atoms/molecules get very fast.
take a configuration that is not too bad, minimize again (longer?)
and see how far you can get.

RP> disabled line with minimization, then RATTLE complained. this is
RP> getting annoying, nothing in comparison to the ease of the water
RP> sphere simulations.

please note that you cut through a water restart and that between
the periodic sides the interactions are in no way in equilibrium.

RP> i have a general question. i am simulating 30 residue peptide and
RP> simulations in water sphere were just perfect for me, but my boss said
RP> that for consistency i need to run PBC of peptide in water box.
RP>
RP> can PBC really so greatly affect simulations? what are the cases when
RP> PBC important? we are interested in just folding the peptide.

yes. in a sphere you have surface tension affecting the pressure
and more importantly a surface dipole across the surface that messes
up your electrostatics. other than for some qualitative tests, i would
be very careful with simulations in a droplet.

there are some codes that use droplet boundary conditions, but they
employ special potentials and random "kicks" on the waters at the
boundary to break the surface dipole. electrostatics are not to be
underestimated when you have regular structures or surfaces...

cheers,
    axel.

RP>
RP>
RP>
RP> On Tue, Jul 29, 2008 at 12:50 AM, Joshua Adelman <jadelman_at_berkeley.edu> wrote:
RP> > I don't think you want to be adding that much 'padding' to the basis
RP> > vectors. Generally I add a fraction of an angstrom beyond my system size. I
RP> > would try decreasing your cellBasisVector's and give it another shot.
RP> >
RP> > Also if you search your error on the namd-l archive, there are several
RP> > discussions with solutions. I would always advise searching that resource
RP> > first since many questions are already answered there.
RP> >
RP> > Josh
RP> >
RP> >
RP> >
RP> > On Jul 28, 2008, at 9:29 PM, Roman Petrenko wrote:
RP> >
RP> >> Dear all,
RP> >> my system size is (min and max)
RP> >> {-16.3659992218 -18.3840007782 -13.1579999924} {17.0270004272
RP> >> 17.5310001373 7.91699981689}
RP> >>
RP> >> the center is
RP> >> 0.253832936287 -0.271096855402 -2.62966871262
RP> >>
RP> >> additionally i add 2.5A in each direction for the basis vectors
RP> >> TO_NAMD:cellBasisVector1 35.892999649 0. 0.
RP> >> TO_NAMD:cellBasisVector2 0. 38.4150009155 0.
RP> >> TO_NAMD:cellBasisVector3 0. 0 23.5749998093
RP> >> TO_NAMD:cellOrigin 0.253832936287 -0.271096855402 -2.62966871262
RP> >> TO_NAMD:wrapAll on
RP> >>
RP> >> now minimization is ok for 100 steps, but the main run (it is NPT
RP> >> simulation) fails no matter what parameter i put for PMEGridSize. i've
RP> >> tried starting from 32 to 90.
RP> >>
RP> >> TO_NAMD:PME yes
RP> >> TO_NAMD:PMEGridSizeX 32
RP> >> TO_NAMD:PMEGridSizeY 32
RP> >> TO_NAMD:PMEGridSizeZ 32
RP> >>
RP> >>
RP> >> every time i get this error:
RP> >> ------------- Processor 0 Exiting: Called CmiAbort ------------
RP> >> Reason: FATAL ERROR: Periodic cell has become too small for original patch
RP> >> grid!
RP> >> Possible solutions are to restart from a recent checkpoint,
RP> >> increase margin, or disable useFlexibleCell for liquid simulation.
RP> >>
RP> >> Charm++ fatal error:
RP> >> FATAL ERROR: Periodic cell has become too small for original patch grid!
RP> >> Possible solutions are to restart from a recent checkpoint,
RP> >> increase margin, or disable useFlexibleCell for liquid simulation.
RP> >> ------------------------------------------------------------------------
RP> >>
RP> >> I can't follow these instructions, because my useFlexibleCell is aready
RP> >> off
RP> >>
RP> >> useGroupPressure yes ;# needed for rigidBonds
RP> >> useFlexibleCell no
RP> >> useConstantArea no
RP> >>
RP> >> Please, help.
RP> >>
RP> >>
RP> >>
RP> >> --
RP> >> Roman Petrenko.
RP> >> Physics Department
RP> >> University of Cincinnati
RP> >>
RP> >
RP> > ------------------------------------------------------------------------------------------------------
RP> > Joshua L. Adelman
RP> > Biophysics Graduate Group Lab: 510.643.2159
RP> > 218 Wellman Hall Fax: 510.642.7428
RP> > University of California, Berkeley http://nature.berkeley.edu/~jadelman
RP> > Berkeley, CA 94720 USA jadelman_at_berkeley.edu
RP> > ------------------------------------------------------------------------------------------------------
RP> >
RP> >
RP> >
RP> >
RP>
RP>
RP>
RP>

-- 
=======================================================================
Axel Kohlmeyer   akohlmey_at_cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:48:05 CST