Comments on parallel-IO

NOTE: The netcdf-hdf mailing list is no longer active.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

I'm still working through the netCDF4 feature list in
http://my.unidata.ucar.edu/content/software/netcdf/netcdf-4/index.html?printer&unidata=47bd4c17e6a6f09cd7f34013dc8c1f6b 

I'm not done reading and thinking, but here are a few comments on the
"Parallel I/O" item, number 6.

One thing that I think needs to be clear is that Parallel I/O will be
supported for netCDF4 HDF5 files, not for netCDF3 files. I think there is also an implicit requirement that all APIs and features 
from regular netCDF4 work as expected in parallel. (This is the case
for HDF5.)

Parallel support has two parts (not counting tuning and mainenance):
  * required API features
  * library features
Interpreting the feature list as a description of tasks, I would note that the parallel support is mostly design work in the first year (making sure 
we don't design something that can't be parallelized), and implemented
in the second year. Notably, we might well release something without
parallel, and might also put in chunking and other features before parallel. In other words, I'm not sure the order of items in the list is necessarily 
chronological.


--
Robert E. McGrath  (HDF Java Team Leader)
National Center for Supercomputing Applications
University of Illinois, Urbana-Champaign
Champaign, Illinois 61820
(217)-333-6549

mcgrath@xxxxxxxxxxxxx
  • 2003 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdf-hdf archives: