[gmx-users] how to set domain decomposition for inhomogeneous systems?
zhongjin
zhongjin1000 at yahoo.com.cn
Sun Sep 4 11:37:05 CEST 2011
Dear GMX user,
I am using GMX4.5.4 (double precision) to simulate electrolyte confined in a CNT with polarizable model. The CNT diameter is 3.0 nm and length is 6.0 nm. The box is 5.0*5.0 *6.0, out of CNT is nothing, that is to say, the system is inhomogeneous，please see it by clicking this link http://photo.163.com/hzj1000/list/#aid=228653982&m=0&page=1
As for inhomogeneous systems,we should set the spatial decomposition with option
-dd.
I use mdrun_d -deffnm nvt -v -nt 8 ,there will be errors.
starting mdrun 'cnt in water'
250000 steps, 500.0 ps.
MDStep= 0/ 0 EPot: 1.12436639e+06, rmsF: 3.00e+03
MDStep= 0/ 1 EPot: 1.10979204e+06, rmsF: 5.29e+02
MDStep= 0/ 2 EPot: 1.10935838e+06, rmsF: 1.22e+02
Warning: 1-4 interaction between 97 and 101 at distance 4.724 which is larger than the 1-4 table size 2.000 nm
These are ignored for the rest of the simulation
This usually means your system is exploding,
if not, you should increase table-extension in your mdp file
or with user tables increase the table size
MDStep= 0/ 3 EPot: nan, rmsF: 7.48e+07
MDStep= 0/ 4 EPot: nan, rmsF: 6.88e+07
MDStep= 0/ 5 EPot: nan, rmsF: 7.03e+07
MDStep= 0/ 6 EPot: nan, rmsF: 7.13e+07
MDStep= 0/ 7 EPot: nan, rmsF: 7.02e+07
MDStep= 0/ 8 EPot: nan, rmsF: 5.91e+07
MDStep= 0/ 9 EPot: nan, rmsF: 5.90e+07
MDStep= 0/10 EPot: nan, rmsF: 5.91e+07
MDStep= 0/11 EPot: nan, rmsF: 5.94e+07
MDStep= 0/12 EPot: nan, rmsF: 5.97e+07
MDStep= 0/13 EPot: nan, rmsF: 6.00e+07
MDStep= 0/14 EPot: nan, rmsF: 5.58e+07
MDStep= 0/15 EPot: nan, rmsF: 5.59e+07
MDStep= 0/16 EPot: nan, rmsF: 5.60e+07
MDStep= 0/17 EPot: nan, rmsF: 5.61e+07
MDStep= 0/18 EPot: nan, rmsF: 5.61e+07
MDStep= 0/19 EPot: nan, rmsF: 5.62e+07
step 0: EM did not converge in 20 iterations, RMS force 121.680
step 0MDStep= 1/ 0 EPot: 1.10511682e+06, rmsF: 1.57e+02
MDStep= 1/ 1 EPot: 1.10517815e+06, rmsF: 3.49e+01
MDStep= 1/ 2 EPot: 1.10515438e+06, rmsF: 8.57e+00
MDStep= 1/ 3 EPot: nan, rmsF: 5.65e+07
MDStep= 1/ 4 EPot: nan, rmsF: 5.29e+07
MDStep= 1/ 5 EPot: nan, rmsF: 5.29e+07
MDStep= 1/ 6 EPot: nan, rmsF: 5.29e+07
MDStep= 1/ 7 EPot: nan, rmsF: 5.29e+07
MDStep= 1/ 8 EPot: nan, rmsF: 5.29e+07
MDStep= 1/ 9 EPot: nan, rmsF: 5.29e+07
MDStep= 1/10 EPot: nan, rmsF: 5.29e+07
MDStep= 1/11 EPot: nan, rmsF: 5.31e+07
MDStep= 1/12 EPot: nan, rmsF: 5.31e+07
MDStep= 1/13 EPot: nan, rmsF: 5.31e+07
MDStep= 1/14 EPot: nan, rmsF: 5.31e+07
MDStep= 1/15 EPot: nan, rmsF: 5.31e+07
MDStep= 1/16 EPot: nan, rmsF: 5.31e+07
MDStep= 1/17 EPot: nan, rmsF: 5.31e+07
MDStep= 1/18 EPot: nan, rmsF: 5.31e+07
MDStep= 1/19 EPot: nan, rmsF: 5.31e+07
step 1: EM did not converge in 20 iterations, RMS force 8.566
Received the INT signal, stopping at the next NS step
MDStep= 2/ 0 EPot: 1.10553389e+06, rmsF: 1.03e+02
MDStep= 2/ 1 EPot: 1.10551468e+06, rmsF: 1.95e+01
MDStep= 2/ 2 EPot: 1.10551473e+06, rmsF: 4.56e+00
MDStep= 2/ 3 EPot: nan, rmsF: 7.55e+07
MDStep= 2/ 4 EPot: nan, rmsF: 7.11e+07
MDStep= 2/ 5 EPot: nan, rmsF: 7.10e+07
MDStep= 2/ 6 EPot: nan, rmsF: 7.09e+07
MDStep= 2/ 7 EPot: nan, rmsF: 7.08e+07
MDStep= 2/ 8 EPot: nan, rmsF: 7.08e+07
MDStep= 2/ 9 EPot: nan, rmsF: 7.07e+07
MDStep= 2/10 EPot: nan, rmsF: 7.07e+07
MDStep= 2/11 EPot: nan, rmsF: 7.06e+07
MDStep= 2/12 EPot: nan, rmsF: 7.06e+07
MDStep= 2/13 EPot: nan, rmsF: 7.06e+07
MDStep= 2/14 EPot: nan, rmsF: 7.06e+07
MDStep= 2/15 EPot: nan, rmsF: 7.06e+07
MDStep= 2/16 EPot: nan, rmsF: 7.06e+07
MDStep= 2/17 EPot: nan, rmsF: 7.06e+07
MDStep= 2/18 EPot: nan, rmsF: 7.06e+07
MDStep= 2/19 EPot: nan, rmsF: 7.06e+07
step 2: EM did not converge in 20 iterations, RMS force 4.560
However, I use mdrun_d -deffnm nvt -v -nt 2, it is OK!
Getting Loaded...
Reading file nvt.tpr, VERSION 4.5.4 (double precision)
Starting 2 threads
Loaded with Money
Making 1D domain decomposition 2 x 1 x 1
Back Off! I just backed up nvt.trr to ./#nvt.trr.2#
Back Off! I just backed up nvt.edr to ./#nvt.edr.2#
starting mdrun 'cnt in water'
250000 steps, 500.0 ps.
MDStep= 0/ 0 EPot: 1.12436639e+06, rmsF: 3.00e+03
MDStep= 0/ 1 EPot: 1.10979204e+06, rmsF: 5.29e+02
MDStep= 0/ 2 EPot: 1.10935838e+06, rmsF: 1.22e+02
MDStep= 0/ 3 EPot: 1.10933788e+06, rmsF: 3.43e+01
MDStep= 0/ 4 EPot: 1.10933636e+06, rmsF: 9.99e+00
MDStep= 0/ 5 EPot: 1.10933623e+06, rmsF: 2.91e+00
MDStep= 0/ 6 EPot: 1.10933622e+06, rmsF: 8.43e-01
MDStep= 0/ 7 EPot: 1.10933622e+06, rmsF: 2.53e-01
MDStep= 0/ 8 EPot: 1.10933622e+06, rmsF: 7.66e-02
MDStep= 0/ 9 EPot: 1.10933622e+06, rmsF: 2.30e-02
MDStep= 0/10 EPot: 1.10933622e+06, rmsF: 7.07e-03
MDStep= 0/11 EPot: 1.10933622e+06, rmsF: 2.16e-03
MDStep= 0/12 EPot: 1.10933622e+06, rmsF: 6.59e-04
And I set domain decompostion by myself instead of by defort.
mdrun_d -deffnm nvt -v -nt 8 -dd 1 1 8
It is OK！
Getting Loaded...
Reading file nvt.tpr, VERSION 4.5.4 (double precision)
Starting 8 threads
Loaded with Money
Making 1D domain decomposition 1 x 1 x 8
Back Off! I just backed up nvt.trr to ./#nvt.trr.3#
Back Off! I just backed up nvt.edr to ./#nvt.edr.3#
starting mdrun 'cnt in water'
250000 steps, 500.0 ps.
MDStep= 0/ 0 EPot: 1.12436639e+06, rmsF: 3.00e+03
MDStep= 0/ 1 EPot: 1.10979204e+06, rmsF: 5.29e+02
MDStep= 0/ 2 EPot: 1.10935838e+06, rmsF: 1.22e+02
MDStep= 0/ 3 EPot: 1.10933788e+06, rmsF: 3.43e+01
MDStep= 0/ 4 EPot: 1.10933636e+06, rmsF: 9.99e+00
MDStep= 0/ 5 EPot: 1.10933623e+06, rmsF: 2.91e+00
MDStep= 0/ 6 EPot: 1.10933622e+06, rmsF: 8.43e-01
MDStep= 0/ 7 EPot: 1.10933622e+06, rmsF: 2.53e-01
MDStep= 0/ 8 EPot: 1.10933622e+06, rmsF: 7.66e-02
MDStep= 0/ 9 EPot: 1.10933622e+06, rmsF: 2.30e-02
MDStep= 0/10 EPot: 1.10933622e+06, rmsF: 7.07e-03
MDStep= 0/11 EPot: 1.10933622e+06, rmsF: 2.16e-03
MDStep= 0/12 EPot: 1.10933622e+06, rmsF: 6.59e-04
step 0MDStep= 1/ 0 EPot: 1.10517946e+06, rmsF: 1.04e+02
MDStep= 1/ 1 EPot: 1.10515795e+06, rmsF: 1.94e+01
MDStep= 1/ 2 EPot: 1.10515825e+06, rmsF: 4.48e+00
MDStep= 1/ 3 EPot: 1.10515825e+06, rmsF: 1.10e+00
MDStep= 1/ 4 EPot: 1.10515824e+06, rmsF: 2.83e-01
MDStep= 1/ 5 EPot: 1.10515824e+06, rmsF: 7.69e-02
MDStep= 1/ 6 EPot: 1.10515824e+06, rmsF: 2.23e-02
MDStep= 1/ 7 EPot: 1.10515824e+06, rmsF: 6.54e-03
MDStep= 1/ 8 EPot: 1.10515824e+06, rmsF: 1.98e-03
I am not so sure what I do is proper, anybody have experence in seting domain decomposition for inhomogeneous systems? Could you give me some advices!
Thanks in advance !
Zhongjin He
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110904/443d4b2d/attachment.html>
More information about the gromacs.org_gmx-users
mailing list