Re: question about HDF5 parallel use of unlimited dimensions...

NOTE: The netcdf-hdf mailing list is no longer active. The list archives are made available for historical reasons.

> "Robert E. McGrath" <mcgrath@xxxxxxxxxxxxx> writes:
> > In general, any update to the file in parallel needs to be done
> > thoughtfully, and generally should be done in batches rather than
> > small increments.
> >
> > Extending by single records is inefficient in all cases, but
> > very costly in parallel, since it updates a global state.
> >
> Does that mean parallel programmers generally don't use unlimited
> dimensions? Or they use them, but batch their extends to avoid this
> problem?

    As Bob says, it is certainly possible to use unlimited dimensions in
parallel, but as you say, I think most people batch their extends.

> Or do they use them, but do something else to prevent this problem.

    Nothing else really occurs to me as a good solution. :-)


> Thanks for helping me understand this!
> Ed
> -- 
> Ed Hartnett  -- ed@xxxxxxxxxxxxxxxx

  • 2005 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdf-hdf archives: