[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #BZE-695706]: Re: [netcdfgroup] please try out the netcdf daily snapshot for opendap, hdf4-reading, hdf5-reading, szip compression, etc.

> Ed Hartnett wrote:
> > Howdy all!
> >
> > Here at NetCDF World Headquarters, we are preparing the 4.1 release of
> > the netCDF C/Fortran libraries. There are many exciting new features in
> > this release, including:
> >
> > * Remote data access with the new built-in opendap client.
> > * Reading of existing HDF4/HDF5 data archives with the netCDF C/Fortran
> >   APIs.
> > * A completely rewritten ncgen, which now fully supports netCDF-4.
> > * Use of parallel-netcdf library to do parallel I/O on netCDF classic
> >   and 64-bit offset format files.
> > * The nc-config script to help build netCDF programs.
> > * UDUNITS now distributed with netCDF.
> > * The ability to compress netCDF-4/HDF5 data with szip
> >   compression. (JUST ADDED!)
> >
> > Now would be a great time for our dedicated user community to download
> > the daily snapshot, let us know that it builds cleanly, and try out
> > some of the new features.
> >
> > Get the latest snapshot here:
> > ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz
> >
> > Build it according to instructions here:
> > http://www.unidata.ucar.edu/software/netcdf/docs_snapshot/netcdf-install.html#Quick-Instructions
> >
> > Report results (good or bad) to the support email address:
> > address@hidden
> >
> > Thanks!
> >
> > Ed
> >
> 
> Here are my experiences with the new szip compression feature so far:
> 
> 1) the configure test doesn't detect szip support in HDF5 properly (it
> fails to find H5Z_can_apply_szip in the lib).  Simply bypassing the test
> worked for me.
> 
> 2) the docs for nc_def_var_szip mention the "bits_per_pixel" parameter,
> but this really should be "bits_per_block" (that's the variable that
> HDF5 is using).  So, it took me a while to figure out how to set this
> parameter from the HDF5 docs until I figured out the name was wrong.
> 
> After getting past those two glitches, it seemed to work fine.
> Compression is faster, but not as good with szip compared to zlib/shuffle.
> 

Thanks Jeff, I fixed both of these.

As you may have noticed on the netCDF mailing list, I've removed the capability 
to create szipped netCDF datasets. The read capability remains (for HDF5 
datasets that have been szipped.

Thanks!

Ed
> 
> 
> --
> Jeffrey S. Whitaker         Phone  : (303)497-6313
> Meteorologist               FAX    : (303)497-6449
> NOAA/OAR/PSD  R/PSD1        Email  : address@hidden
> 325 Broadway                Office : Skaggs Research Cntr 1D-113
> Boulder, CO, USA 80303-3328 Web    : http://tinyurl.com/5telg
> 
> 


Ticket Details
===================
Ticket ID: BZE-695706
Department: Support netCDF
Priority: Critical
Status: Closed