Hello,
As professor Case said, there is no error message here. How long does it
stay this way? I would suggest trying to do only a very tiny number of
steps and see what happens (for instance, try 5 steps printing information
every step). Does the job end? If so, what does the stderr stream report?
Do the parallel tests pass? (You should do this all with very small
processor counts, i.e. 1 or 2 as Professor Case suggested). If you do 1
processor, still use sander.MPI, just do something like "mpirun -np 1
sander.MPI" so that the MPI executable is still run.
All the best,
Jason
On Fri, Aug 20, 2010 at 10:31 AM, George Tzotzos <gtzotzos.me.com> wrote:
>
> Thanks for the prompt reply. It's the first time I'm running sander.MPI on
> this machine. The machine has passed all tests successfully and on numerous
> earlier occasions I've run pmemd.MPI on 16 and 32 processors successfully.
>
> amber11-serial produces <mdinfo>, <min.out> and <min.rst> files.
>
> The mdinfo is shown below:
>
> NSTEP ENERGY RMS GMAX NAME NUMBER
> 614 -6.6756E+04 1.3575E+01 3.8625E+01 O 15978
>
> BOND = 29.7887 ANGLE = 288.6140 DIHED =
> 1342.7352
> VDWAALS = 3417.4085 EEL = -80174.8566 HBOND =
> 0.0000
> 1-4 VDW = 542.7587 1-4 EEL = 7780.4877 RESTRAINT =
> 16.7276
> EAMBER = -66773.0640
>
> amber11-parallel on 2 processors (same as on 16) produces an empty <mdinfo>
> and no <min.rst> files. The <min.out> file is as the one sent before. Its
> results section is as below:
>
> 4. RESULTS
>
> --------------------------------------------------------------------------------
>
> ---------------------------------------------------
> APPROXIMATING switch and d/dx switch using CUBIC SPLINE INTERPOLATION
> using 5000.0 points per unit in tabled values
> TESTING RELATIVE ERROR over r ranging from 0.0 to cutoff
> | CHECK switch(x): max rel err = 0.2738E-14 at 2.422500
> | CHECK d/dx switch(x): max rel err = 0.8314E-11 at 2.736960
> On Aug 20, 2010, at 4:03 PM, case wrote:
>
> > On Fri, Aug 20, 2010, George Tzotzos wrote:
> >>
> >> In amber11_serial there's seems to be no problem.
> >>
> >> Yet running the same input files on a 16 processor machine, I get the
> >> error message below.
> >
> > ??? I don't see any error message. Does the job just "hang"?
> >
> > Do the parallel tests work (in particular, things like
> dhfr/Run.dhfr.min?)
> > If so, try to see the difference between what you are doing and test
> case.
> > You might try fewer processors (even 1) to narrow down the problem.
> >
> > ....dac
> >
> >
> > _______________________________________________
> > 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
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 20 2010 - 09:00:03 PDT
