[gmx-users] Re: equilibrium stage runs for too long

[Justin Lemkul]

On 8/24/12 6:01 AM, Clare wrote:
> Dear Vitaly,
>
> Thank you so much for your suggestion!!
> I've run the simultion using one core as you suggested, and I can see two
> outputs on step 0 and step 100 now ( shown on the bottom), but it still took
> a while to only generate these two outputs.
> Maybe it is indeed the problem of scaling as mentioned by Justin?
>
> There are: 34001 Atoms

The reason that it is slow now is that you have 34001 atoms being calculated on 
one core.  What that demonstrates is that your simulation is stable and likely 
did not crash when run in parallel.  My previous comments about scaling issues 
were simply a guess based on incomplete information.  Your log file from the MPI 
run indicated the job never even started.  That's not an issue of scaling; it's 
an issue of functionality.  It looks to me like either mdrun_mpi or your MPI 
implementation is broken or not being invoked properly.  Try the test case I 
suggested before, something that doesn't involve Gromacs.

-Justin

> Charge group distribution at step 0: 1953 1973 2103 2032 2044 2072 2063 1967
> Grid: 3 x 7 x 18 cells
>
> Constraining the starting coordinates (step 0)
>
> Constraining the coordinates at t0-dt (step 0)
> RMS relative constraint deviation after constraining: 1.99e-05
> Initial temperature: 300.057 K
>
> Started mdrun on node 0 Fri Aug 24 10:37:20 2012
>
>             Step           Time         Lambda
>                0        0.00000        0.00000
>
>     Energies (kJ/mol)
>            Angle    Proper Dih. Ryckaert-Bell.  Improper Dih.          LJ-14
>      6.66008e+03    3.06081e+03    7.94034e+02    2.55031e+02    8.61710e+02
>       Coulomb-14        LJ (SR)  Disper. corr.   Coulomb (SR)   Coul. recip.
>      2.40538e+03    5.11091e+04   -3.00143e+03   -4.49637e+05   -8.11103e+04
>   Position Rest.      Potential    Kinetic En.   Total Energy  Conserved En.
>      5.53187e+01   -4.68547e+05    8.57672e+04   -3.82780e+05   -3.82780e+05
>      Temperature Pres. DC (bar) Pressure (bar)   Constr. rmsd
>      3.00450e+02   -9.73436e+01   -2.97627e+03    3.74608e-05
>
> DD  step 4 load imb.: force  6.0%
>
> DD  step 99 load imb.: force  6.1%
>
>             Step           Time         Lambda
>              100        0.20000        0.00000
>
>     Energies (kJ/mol)
>            Angle    Proper Dih. Ryckaert-Bell.  Improper Dih.          LJ-14
>      1.10274e+04    2.82478e+03    1.23525e+03    2.24466e+02    1.54991e+03
>       Coulomb-14        LJ (SR)  Disper. corr.   Coulomb (SR)   Coul. recip.
>      4.67051e+03    6.34091e+04   -3.00143e+03   -4.31792e+05   -8.19081e+04
>   Position Rest.      Potential    Kinetic En.   Total Energy  Conserved En.
>      1.44619e+04   -4.17298e+05    7.13854e+04   -3.45913e+05   -3.82303e+05
>      Temperature Pres. DC (bar) Pressure (bar)   Constr. rmsd
>      2.50069e+02   -9.73436e+01   -3.79866e+02    3.45085e-06
>
>
>
>
>
> --
> View this message in context: http://gromacs.5086.n6.nabble.com/equilibrium-stage-runs-for-too-long-tp5000461p5000483.html
> Sent from the GROMACS Users Forum mailing list archive at Nabble.com.
>

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================