mirror/netcdf-c
2
0
Fork 0
mirror of https://github.com/Unidata/netcdf-c.git synced 2025-01-30 16:10:44 +08:00
Code Issues Packages Projects Releases Wiki Activity
2bf233c9d5
netcdf-c/libsrc4
History
M. Scot Breitenfeld 2bf233c9d5 This patch changes the algorithm for determining the extended size of a dataset in parallel to pass a variable of type unsigned long long to MPI_Allreduce. Despite the comment in the code on this line (removed in this patch), the current usage is not correct. For example, consider if process 0 has an extend size of 2^32 (0x100000000) and process 2 has an extend size of 1 (0x1). The current algorithm will compute the max of each 4 byte segment then combine these into an 8 byte number, yielding a max of (2^32)+1 (0x100000001), when it should simply be 2^32. 
N. Fortner
2016-06-21 13:04:15 -05:00
..
CMakeLists.txt Add provenance info for netcdf-4 files. 2016-05-07 14:32:07 -06:00
error4.c merged trunk up to revision 75, move v2 API 2010-06-07 11:12:14 +00:00
Makefile.am Add provenance info for netcdf-4 files. 2016-05-07 14:32:07 -06:00
nc3stub.c Refactored read_scale(), memio_new(), var_create_dataset() and makespecial() 2013-12-08 03:29:26 -06:00
nc3stub.h Quincey's fixes for NCF-250, netCDF-4 parallel independent access with 2013-07-08 21:31:13 +00:00
nc4attr.c HDF5 is generating unnecessary error messages when netcdf4 logging is enabled 2016-05-27 10:08:01 -06:00
nc4dim.c Finally get around to removing all that 2016-05-14 22:31:41 -06:00
nc4dispatch.c merge-squash 2015-10-09 10:12:11 -06:00
nc4dispatch.h [NCF-273]/HZY-708311 2013-12-22 12:53:20 -07:00
nc4file.c The name hash for hdf4 variables was 2016-06-01 15:20:36 -06:00
nc4grp.c Modified a number of tests 2014-10-04 15:59:16 -06:00
nc4hdf.c This patch changes the algorithm for determining the extended size of a dataset in parallel to pass a variable of type unsigned long long to MPI_Allreduce. Despite the comment in the code on this line (removed in this patch), the current usage is not correct. For example, consider if process 0 has an extend size of 2^32 (0x100000000) and process 2 has an extend size of 1 (0x1). The current algorithm will compute the max of each 4 byte segment then combine these into an 8 byte number, yielding a max of (2^32)+1 (0x100000001), when it should simply be 2^32. 2016-06-21 13:04:15 -05:00
nc4info.c HDF5 is generating unnecessary error messages when netcdf4 logging is enabled 2016-05-27 10:08:01 -06:00
nc4internal.c HDF5 is generating unnecessary error messages when netcdf4 logging is enabled 2016-05-27 10:08:01 -06:00
nc4type.c Fixed an issue preventing compilation with hdf4 support with Visual Studio. 2014-08-28 18:14:14 -06:00
nc4var.c Finally get around to removing all that 2016-05-14 22:31:41 -06:00
ncfunc.c Finally get around to removing all that 2016-05-14 22:31:41 -06:00
stub3.c dispatch.h->ncdispatch.h to avoid visualstudio filename conflict 2010-08-05 02:44:59 +00:00
stub4.c fixed some missing config.h problems 2011-09-26 10:14:55 +00:00