[CPMD-list] (no subject)

[Axel Kohlmeyer]

On Tue, 8 May 2007, qfzhang wrote:

QZ> 
QZ> Hi,
QZ>    What do you mean _any_parallel job? It's no problem to run other parallel job
QZ> s, even CP-PIMD with more than one processor groups.

well, it was not clear from your report and the error message, memory 
usage pattern (large address space, little resident set size) would 
also match a miscompiled parallel binary (pointer size mixup).

QZ>    And another problem is that when I do constrained  CP-PIMD calculations, only
QZ>  one processor group method works. Do they have the same cause? 

i don't know the PIMD part in CPMD very well, but it is not impossible
that you may run into bugs in the code. the PIMD code in the recent
releases has not been thoroughly tested or updated in a long time, 
whereas most of the other parts of the code have been improved/changed. 
so it is quite possible that some of the PIMD functionality needs to 
be recovered.

as for using constraints in PIMD, you may want to search through the
mailing list archives, as i seem to remember vaguely, that there was 
a discussion on that quite a while ago...

cheers,
   axel.

QZ>   
QZ>    Best wishes
QZ>                                                        Qianfan Zhang
QZ>                                                        2007-05-08
QZ> 
QZ> Axel Kohlmeyer д:
QZ> 
QZ> > On Tue, 8 May 2007, qfzhang wrote:
QZ> > 
QZ> > 
QZ> > can you run _any_ parallel job with that executable?
QZ> > 
QZ> > axel.
QZ> > 
QZ> > QZ> Hi,dear all!
QZ> >    I want to use BO-PIMD for calculations, but I find that the job will dead a
QZ> 
QZ> ft
QZ> > er reading RESTART files, and when I specify the parameter PROCESSOR GROUPS,an
QZ> 
QZ> d 
QZ> > let it equal 1, the job works,and the only way to make it work.But the efficie
QZ> 
QZ> nc
QZ> > y is not high.Why can\\'t I use more than one processor groups? the input file
QZ>  
QZ> is 
QZ> > as below.
QZ> >   &CPMD
QZ> >   PATH INTEGRALS
QZ> >   MOLECULAR DYNAMICS BO
QZ> >   RESTART WAVEFUNCTION COORDINATES LATEST
QZ> >   CONVERGENCE ORBITALS
QZ> >    1.0e-4
QZ> >   ODIIS
QZ> >    15
QZ> >    CONVERGENCE RELAX
QZ> >    20
QZ> >    TEMPERATURE
QZ> >     300.d0
QZ> >    TEMPCONTROL IONS
QZ> >    300 20
QZ> >    MAXSTEP
QZ> >     9000
QZ> >    TIMESTEP
QZ> >     15
QZ> > &END
QZ> > 
QZ> > &DFT
QZ> >   NEWCODE
QZ> >   FUNCTIONAL BLYP
QZ> > &END
QZ> > 
QZ> > &SYSTEM
QZ> >   ANGSTROM
QZ> >   SYMMETRY
QZ> >   1
QZ> >   CELL
QZ> >   4.13 1.0 1.0  0  0  0
QZ> >   SCALE S=1
QZ> >   CUTOFF
QZ> >   80.0
QZ> >   CHARGE
QZ> >   1.0
QZ> > &END
QZ> > 
QZ> > &ATOMS
QZ> > ........
QZ> > ........
QZ> > &END
QZ> > 
QZ> > &PIMD
QZ> >    CENTROID DYNAMICS
QZ> >    TROTTER DIMENSION
QZ> >      16
QZ> >    NORMAL MODES
QZ> >      25.0
QZ> >    FACMASS
QZ> >      1.0
QZ> >    PRINT LEVEL
QZ> >      1
QZ> >    PROCESSOR GROUPS
QZ> >     4
QZ> > &END
QZ> >    And the final part of the output file is as below
QZ> >  INITIALIZATION TIME:                                1.00 SECONDS
QZ> > 
QZ> >  ***   PI_MDPT| SIZE OF THE PROGRAM IS   34012/2054232 kBYTES ***
QZ> >  RV30| WARNING! NO WAVEFUNCTION VELOCITIES
QZ> >  RESTART INFORMATION READ ON FILE                   ./RESTART_1.1
QZ> >  RESTART INFORMATION READ ON FILE                   ./RESTART_2.1
QZ> >  RESTART INFORMATION READ ON FILE                   ./RESTART_3.1
QZ> >  RESTART INFORMATION READ ON FILE                   ./RESTART_4.1
QZ> > p0_9445:  p4_error: interrupt SIGx: 6
QZ> > rm_l_2_1817: (25.589844) net_send: could not write to fd=5, errno = 104
QZ> > 
QZ> > 
QZ> >   Best wishes
QZ> >                                             Qianfan Zhang
QZ> >                                              2007-05-08
QZ> > 
QZ> > 
QZ> > -- 
QZ> > =======================================================================
QZ> > Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
QZ> >    Center for Molecular Modeling   --   University of Pennsylvania
QZ> > Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
QZ> > tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
QZ> > =======================================================================
QZ> > If you make something idiot-proof, the universe creates a better idiot.
QZ> > 
QZ> 
QZ> 
QZ> 

-- 
=======================================================================
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.