We have continued to shake out bugs in the "big variable" file format. (See NewFileFormatDefinition for technical details and NewFileFormatCode for usage. Thank Kui Gao from Northwestern University for doing a lot of the work). Thanks to our friends at PNNL for giving the "big variable" format a good shakedown before the final release.

We introduced a new optimization in parallel-netcdf with this release. If the hint "striping_unit" is set, then pnetcdf will align the start of non-record variables to a multiple of that value. See StripingUnitHint for more information.

The 'ncgen' utility can now create "big variable" (CDF-5) files, in addition to the older "big file" (CDF-2) files.

There is now a hand-written pnetcdf.F90 module for F90 codes, based on the one written by Annette Koontz of PNL. Please consider this as an "early feedback" version and not a hard gaurantee of a fixed F90 API (though we don't expect it to change drastically if at all)

Kui Gao has been busy! He implemented "write combining" via the nonblocking interfaces. An application can now post multiple writes or reads to different variables and the pnetcdf library will combine those operations into a single I/O request. See CombiningOperations for some more documentation.

We did some other stuff, too. Still need to go through the svn logs.

Download URLs:

• bzip2ded tarball
• gzipped tarball:

SHA-1 Checksums:

SVN tag: