Changes between Version 5 and Version 6 of ReleaseNotes-1.1.0

Show
Ignore:
Timestamp:
10/27/09 13:15:59 (4 weeks ago)
Author:
robl (IP: 205.208.56.252)
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • ReleaseNotes-1.1.0

    v5 v6  
    1 We have continued to shake out bugs in the "big variable" file format.  (See 
    2 NewFileFormatDefinition for technical details and NewFileFormatCode for usage. 
    3 Thank Kui Gao from Northwestern University for doing a lot of the work). 
    4 Thanks to our friends at PNNL for giving the "big variable" format a good 
    5 shakedown before the final release. 
     1= Parallel-NetCDF 1.1.0 = 
    62 
    7 We introduced a new optimization in parallel-netcdf with this release.  If the 
    8 hint "striping_unit" is set, then pnetcdf will align the start of non-record 
    9 variables to a multiple of that value.  See StripingUnitHint for more 
    10 information. 
     3We released version 1.1.0 on October 27th, 2009.  This release marked a big 
     4milestone for pnetcdf and included several new features along with the usual 
     5bug fixes. 
    116 
    12 The 'ncgen' utility can now create "big variable" (CDF-5)  files, in addition to the older "big file" (CDF-2) files. 
     7== New Features ==  
    138 
    14 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) 
    15  
    16 Kui Gao has been busy! He implemented "write combining" via the nonblocking interfaces.  An application can 
    17 now post multiple writes or reads to different variables and the pnetcdf library will combine those operations into a 
    18 single I/O request.  See CombiningOperations for some more documentation. 
    19  
    20 We did some other stuff, too.  Still need to go through the svn logs. 
     9  * A new file format, CDF-5, is introduced. This format allows defining large 
     10   array variables with more than 2^32^ elements. Please see 
     11   NewFileFormatDefinition for technical details and NewFileFormatCode for 
     12   usage.  We have shaken out a lot of bugs since the pre-release. 
     13  * We introduced a new optimization in parallel-netcdf with this release.  If 
     14   the hint "striping_unit" is set, then pnetcdf will align the start of 
     15   non-record variables to a multiple of that value.  See StripingUnitHint 
     16   for more information. 
     17  * Parallel-netCDF now contains a new set of APIs for reading/writing 
     18   multiple variables. The existing asynchronous APIs have also been improved 
     19   to enable combination of multiple variable access into fewer I/O requests. 
     20   Please see CombiningOperations for some further information. 
     21  * There is now a simple pnetcdf.F90 module for F90 codes, based on the one 
     22    written by Annette Koontz of PNL.  Please consider this as an "early 
     23    feedback" version and not a hard gaurantee of a fixed F90 API (though we 
     24    don't expect it to change drastically if at all) 
     25  * The 'ncgen' utility can now create "big variable" (CDF-5)  files, in 
     26   addition to the older "big file" (CDF-2) files. 
    2127 
    2228Download URLs: