ok- that makes more sense. if you stop and start, there is limited precision
to the restart files so these are not expected to match.
the way you initially explained it I thought the jobs were the same- but i
suspected that you would not be able to write that frequently and have a
long run.
of course many short runs are not the same as 1 long one. starting and
stopping frequently can also introduce many kinds of artifacts, particularly
with the Langevin thermostat, there is literature on that.
your script below doesn't make sense to me- what are you trying to do? why
convert to charmm format?
2010/4/13 Jianing <song_jianing_hi.126.com>
> Hi Simmerling,
> Actually, I haven't diffed the mdout files.
> but maybe I have found the reason for this surprising phenomenon, and I
> need you to give a juge on this.
> you know, writing every 4fs for 3.5ns is a very large mdcrd file,
> especially when the trajectory is not a binary file, so I just cut it into
> hunderds of pieces. That is, when sander finishes every 10ps interval,it
> will output md1.rst, then, when do the next 10ps, the initial coordinates is
> from md1.rst. and this cycle lasts until 3.5ns is reached. Although this is
> proceeded by computer automatically, this may also bright in errors, because
> the data accuracy is different between calculating and reading data in
> sander, which I didn't know before.
> if this is the truth, I'm so sorry for disturbing you so much.
> Thank you very much.
> Jianing
>
>
>
>
> $B:_(B2010-04-12 22:02:01$B!$(B"Carlos Simmerling" <carlos.simmerling.gmail.com> $B<LF;!'(B
> >have you diffed the mdout files?
> >
> >2010/4/12 Jianing <song_jianing_hi.126.com>
> >
> >> when the coordinates are saved every 4fs, the program sander just exits
> >> automatically at 3.5ns for the dynamic run. and the XRD structure
> bacomes
> >> unstable at 2ns. The time is not very long, so this also surprises me. I
> >> change to pmemd and get the same results.
> >> you may say there is something wrong with my operation, and the
> following
> >> is my executing file to demenstrate that the whole process is proceeded
> by
> >> computer automatically, and I don't change anything during the
> simulation.
> >>
> >> run.sh:
> >>
> >> for i in `seq 1 500`
> >>
> >> do
> >>
> >> let "j=i-1"
> >>
> >> mpirun -n 8 pmemd -O -i md.in -o md${i}.out -p vilppcxrd.prm -c
> >> md${j}.rst -r md${i}.rst -x md${i}.mdcrd </dev/null
> >>
> >> echo "
> >>
> >> trajin md${i}.mdcrd
> >>
> >> trajout md${i}.dcd charmm " > p.in
> >>
> >> ptraj vilppcxrd.prm p.in
> >>
> >> rm p.in
> >>
> >> rm md${i}.mdcrd
> >>
> >> done
> >>
> >> what I just have done is prepare the input file md.in and the initial
> >> parmtop file and coordinates file vilppcxrd.prm and md0.rst.
> >>
> >> Thanks
> >>
> >> Jianing
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >> $B:_(B2010-04-12 20:13:31$B!$(B"Carlos Simmerling" <carlos.simmerling.gmail.com>
> $B<LF;!'(B
> >> >no- what I meant is what is the length of simulation you are comparing?
> >> how
> >> >long before the XRD structure becomes unstable?
> >> >
> >> >2010/4/12 Jianing <song_jianing_hi.126.com>
> >> >
> >> >> Thank you very much for your reply.
> >> >> yes, I'm sure using the same Amber binary, the same machine, the same
> >> >> number of CPU and comparing the same time interval. The results shock
> >> me, so
> >> >> I have checked everything that may affect the simulation.
> >> >>
> >> >>
> >> >> >writing every 4fs is very often, and would
> >> >> >probably give very laerge output files for a simulatino long enough
> to
> >> >> lose
> >> >> >the XRD structure
> >> >>
> >> >> what is the extrat meaning of this? Do you mean that writing the
> >> >> coordinates too often is equil to run MD simulation for a long time
> that
> >> may
> >> >> cause to lose XRD structure.
> >> >>
> >> >> Thanks in advance.
> >> >>
> >> >> Jianing
> >> >>
> >> >>
> >> >>
> >> >> $B:_(B2010-04-12 19:09:45$B!$(B"Carlos Simmerling" <
> carlos.simmerling.gmail.com>
> >> $B<LF;!'(B
> >> >> >the frequency of saving does not change the simulation. Are you sure
> >> that
> >> >> >you ran them using the same Amber binary, on the same machine, using
> >> the
> >> >> >same # of CPus, etc? many things can cause trajectories to diverge.
> >> Have
> >> >> you
> >> >> >diffed the mndout files? are you sure that you are comparing the
> same
> >> time
> >> >> >writing every 4fs is very often, and would
> >> >> >probably give very laerge output files for a simulatino long enough
> to
> >> >> lose
> >> >> >the XRD structure.
> >> >> >
> >> >> >2010/4/12 Nancy <song_jianing_hi.126.com>
> >> >> >
> >> >> >> Dear all,
> >> >> >>
> >> >> >> Recently I'm involved in performing MD simulation using AMBER.
> >> >> >> At the beginning, the atom coordinates are saved to mdcrd file in
> >> every
> >> >> >> 4fs(ntrw=2) time interval. Then for comparison, I save the atom
> >> >> coordinates
> >> >> >> into mdcrd file every 10ps(ntrw=5000). Other parts of the two MD
> >> >> simulations
> >> >> >> are all the same, and the only difference is the frequency of
> saving
> >> >> atom
> >> >> >> coordinates. But the conformation between the two simulation give
> >> huge
> >> >> >> difference. That from former protocol is far from the XRD
> structure
> >> at
> >> >> the
> >> >> >> end. But that from the latter is similar to the XRD structure.
> >> >> >>
> >> >> >> This results shocks me, because the frequency of saving atom
> >> coordinates
> >> >> >> shouldn't affect to MD simuation. I've checked all the input
> files,
> >> they
> >> >> are
> >> >> >> the same, except options controlling the output such as 'ntwx'.
> The
> >> >> program
> >> >> >> I use is sander in AMBER10, AMBER03 is the force field by which
> I've
> >> >> adopted
> >> >> >> to perform simulation. My system is villin headpiece. Furthermore,
> >> >> explicit
> >> >> >> solvation model TIP3P is used.
> >> >> >>
> >> >> >> The following is the two input file:
> >> >> >>
> >> >> >> ---The frequently-output input-------------
> >> >> >> &cntrl
> >> >> >> imin = 0,
> >> >> >> irest = 1,
> >> >> >> ntx = 5,
> >> >> >> ntb = 1,
> >> >> >> cut = 12.0,
> >> >> >> ntc = 2,
> >> >> >> ntf = 2,
> >> >> >> tempi = 300.0,
> >> >> >> temp0 = 300.0,
> >> >> >> ntt = 3,
> >> >> >> gamma_ln = 1.0,
> >> >> >> nstlim = 5000, dt = 0.002
> >> >> >> ntpr = 2, ntwx = 2, ntwr = 2
> >> >> >> /
> >> >> >> ----------------------------------------
> >> >> >> ---The not-so-frequently-output input---
> >> >> >> &cntrl
> >> >> >> imin = 0,
> >> >> >> irest = 1,
> >> >> >> ntx = 5,
> >> >> >> ntb = 1,
> >> >> >> cut = 12.0,
> >> >> >> ntc = 2,
> >> >> >> ntf = 2,
> >> >> >> tempi = 300.0,
> >> >> >> temp0 = 300.0,
> >> >> >> ntt = 3,
> >> >> >> gamma_ln = 1.0,
> >> >> >> nstlim = 2500000, dt = 0.002
> >> >> >> ntpr = 5000, ntwx =5000 , ntwr = 5000
> >> >> >> /
> >> >> >> ----------------------------------------
> >> >> >> can someone give me some advice and help me out of the trouble?
> >> >> >> Thank you very much in advance.
> >> >> >> Jianing
> >> >> >> _______________________________________________
> >> >> >> AMBER mailing list
> >> >> >> AMBER.ambermd.org
> >> >> >> http://lists.ambermd.org/mailman/listinfo/amber
> >> >> >>
> >> >> >_______________________________________________
> >> >> >AMBER mailing list
> >> >> >AMBER.ambermd.org
> >> >> >http://lists.ambermd.org/mailman/listinfo/amber
> >> >>
> >> >> _______________________________________________
> >> >> AMBER mailing list
> >> >> AMBER.ambermd.org
> >> >> http://lists.ambermd.org/mailman/listinfo/amber
> >> >>
> >> >>
> >> >_______________________________________________
> >> >AMBER mailing list
> >> >AMBER.ambermd.org
> >> >http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >>
> >_______________________________________________
> >AMBER mailing list
> >AMBER.ambermd.org
> >http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 13 2010 - 05:00:03 PDT
