H5Part development and discussion

[John Shalf <JShalf AT lbl.gov>]

On Dec 1, 2006, at 3:16 PM, Kurt Stockinger wrote:

> John Shalf wrote:
> > Hi Kurt,
> > I think SetView is inclusive of the starting particle address but
> > exclusive of the ending address.  I can't remember the rationale for
> > that, but I think that is the case.
> Hi John,
>
> You are right on the behavior of SetView. The problem seems with the
> return value of H5GetNumOfParticles:

But if you agree with my semantics that I laid out above for SetView,  
then the return value offered by GetNumOfParticles() is correct.
        After SetView(10,14): steps= 5  datasets= 7 particles= 4
4 is the correct number of particles to return because it is  
inclusive of 10 and exclusive of particle #14.

-john

> H5Part.c:
> h5part_int64_t
> _H5Part_get_num_particles (
> ..
>     if ( H5PartHasView ( f ) ) {
>         nparticles = H5Sget_select_npoints ( space_id );
>     }
>
> nparticles is off by 1.
>
> Kurt
> > -john
> >
> > On Dec 1, 2006, at 2:23 PM, Kurt Stockinger wrote:
> >
> > > Hi,
> > >
> > > I just noticed that H5PartGetNumParticles() seems to return the  
> > > wrong
> > > number of particles. It's always off by one, i.e. n-1 rather than n.
> > > I've tested the program H5PartTest with 4 processes and 5 particles
> > > each.
> > >
> > > Thanks,
> > > Kurt
> > >
> > > trunk/test> mpirun -np 4 ./H5PartTest
> > > Proc[0] Writing timestep 0 file=67108864
> > > ..
> > > Proc[3] Writing timestep 4 file=67108864
> > > AllDone p[1]
> > > ..
> > > After SetView(0,4): steps= 5  datasets= 7 particles= 4
> > > After SetView(15,19): steps= 5  datasets= 7 particles= 4
> > > nParticles: 4
> > > nParticles: 4
> > > After SetView(10,14): steps= 5  datasets= 7 particles= 4
> > > nParticles: 4
> > > After SetView(5,9): steps= 5  datasets= 7 particles= 4
> > > nParticles: 4
> > > proc[2]:  done
> > > proc[1]:  done
> > > proc[3]:  done
> > > proc[0]:  done
> > >
> > >
> > > trunk/test> h5dump parttest.h5
> > > HDF5 "parttest.h5" {
> > > GROUP "/" {
> > >    GROUP "Step#0" {
> > >       DATASET "id" {
> > >          DATATYPE  H5T_STD_I64LE
> > >          DATASPACE  SIMPLE { ( 20 ) / ( 20 ) }
> > >          DATA {
> > >          (0): 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
> > > 16, 17,
> > >          (18): 18, 19
> > >          }
> > >       }
> > >       DATASET "px" {
> > >          DATATYPE  H5T_IEEE_F64LE
> > >          DATASPACE  SIMPLE { ( 20 ) / ( 20 ) }
> > >          DATA {
> > >          (0): 0, 1, 2, 3, 4, 10, 11, 12, 13, 14, 20, 21, 22, 23, 24,
> > > 30, 31,
> > >          (17): 32, 33, 34
> > >          }
> > >       }
> > >
> > >
> > > --Kurt Stockinger
> > > Computational Research Division
> > > Lawrence Berkeley National Laboratory
> > > Mail Stop 50B-3238, 1 Cyclotron Road
> > > Berkeley, California 94720, USA
> > >
> > > Tel: +1 (510) 486 5208, Fax: +1 (510) 486 4004
> > > email: KStockinger AT lbl.gov
> > > http://sdm.lbl.gov/kurts/
> > >
> > > _______________________________________________
> > > H5Part mailing list
> > > H5Part AT lists.psi.ch
> > > https://lists.web.psi.ch/mailman/listinfo/h5part
>
>
> --
> Kurt Stockinger
> Computational Research Division
> Lawrence Berkeley National Laboratory
> Mail Stop 50B-3238, 1 Cyclotron Road
> Berkeley, California 94720, USA
>
> Tel: +1 (510) 486 5208, Fax: +1 (510) 486 4004
> email: KStockinger AT lbl.gov
> http://sdm.lbl.gov/kurts/
>
> _______________________________________________
> H5Part mailing list
> H5Part AT lists.psi.ch
> https://lists.web.psi.ch/mailman/listinfo/h5part