repeated output entries

sersunzo Pitago... at gmail.com
Tue Nov 18 13:24:08 UTC 2008


Teo,

I'm not sure I understand. The following file (ARCH=Linux-x86-64-
intel.popt) was used for compilation

CC       = mpicc
CPP      =
FC       = mpif90
LD       = mpif90
AR       = ar -r
DFLAGS   = -D__INTEL -D__FFTSG -D__parallel -D__BLACS -D__SCALAPACK -
D__FFTW3
CPPFLAGS =
FCFLAGS  = $(DFLAGS) -O3 -xW -funroll-loops -fpp -free
LDFLAGS  = $(FCFLAGS)
LIBS     = /share/apps/scalapack/libscalapack.a \
            /share/apps/scalapack/blacs.a \
            /share/apps/scalapack/blacsC.a \
            /share/apps/scalapack/blacsF77.a \
            /share/apps/scalapack/blacs.a \
            /share/apps/scalapack/libreflapack.a \
            /share/apps/acml4.1.0/ifort64/lib/libacml.a\
            /share/apps/acml4.1.0/ifort64/lib/libacml_mv.a \
            /usr/local/lib/libfftw3.a

OBJECTS_ARCHITECTURE = machine_intel.o

Is it not for .popt?

Alessio

On Nov 17, 12:38 pm, Teodoro Laino <teodor... at gmail.com> wrote:
> Alessio,
>
> that's a typical case when you submit sopt with mpirun.
> please, compile for popt.
> teo
>
>
>
> sersunzo wrote:
> > Dear all,
>
> > I’m a new user to CP2K. I’m trying to run some exercises repeating
> > some cases proposed at a tutorial I attended in Lyon this May.
>
> > When I run a MD simulation on a AMD-Opteron (2.2 GHz) cluster I
> > noticed that the text in output files is repeated according to the
> > number of processor I used.
>
> > For instance, the .ener file with 1 processor is
>
> > Step       Time[fs]        Kin.[a.u.]      …                    UsedTime[s]
> > 0  0       0.012825602     …     0
> > 1  0.5     0.012483648     …     15.3102
> > 2  1       0.011524168     …     6.7489
>
> > with 2 processors
>
> > Step       Time[fs]        Kin.[a.u.]      …                    UsedTime[s]
> > 0  0       0.012825602     …     0
> > 1  0.5     0.012483669     …     26.1512
> > 1  0.5     0.012483669     …     26.3826
> > 2  1       0.011524142     …     6.7792
> > 2  1       0.011524142     …     6.7637
> > …etc.
>
> > with four
>
> > Step       Time[fs]        Kin.[a.u.]      …                    UsedTime[s]
> > 0  0       0.012825602     …     0
> > 1  0.5     0.012483669     …     52.7459
> > 1  0.5     0.012483669     …     53.7512
> > 1  0.5     0.012483669     …     57.6845
> > 1  0.5     0.012483669     …     57.689
> > 2  1       0.011524142     …     13.4656
> > 2  1       0.011524142     …     14.3997
> > 2  1       0.011524142     …     13.8007
> > 2  1       0.011524142     …     14.1252
> > 3  1.5     0.010195152     …     17.2202
> > 3  1.5     0.010195152     …     17.6331
> > 3  1.5     0.010195152     …     17.1964
> > 3  1.5     0.010195152     …     17.4447
> > 4  2       0.008830685     …     13.4772
> > 4  2       0.008830685     …     14.5027
> > 4  2       0.008830685     …     14.0316
> > 4  2       0.008830685     …     13.9997
> > …etc.
>
> > It looks like each processor is working by its own and the calculation
> > time increases with the number of processors. I supposed there is a
> > compilation issue (arch=Linux-x86-64-intel). I just wondering if
> > somebody else had this problem before.
>
> > Thank in advance
> > Alessio- Hide quoted text -
>
> - Show quoted text -


More information about the CP2K-user mailing list