-
Notifications
You must be signed in to change notification settings - Fork 22
/
README
97 lines (79 loc) · 4.63 KB
/
README
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
PnetCDF is a high-performance parallel I/O library for accessing Unidata's
NetCDF files in classic formats, specifically the CDF 1, 2, and 5 formats.
Although NetCDF supports parallel I/O starting from version 4, it is built on
top of HDF5 and thus its parallel feature requires files in HDF5 format.
PnetCDF is currently the only option for performing parallel I/O on files in
classic formats.
Starting from version 1.1.0, PnetCDF supports the CDF-5 file format, an
extension of CDF-2, that supports more atomic data types and allows users to
define large dimensions, attributes, and variables (>2B array elements).
NetCDF gives scientific programmers a space-efficient and portable means for
storing data. By making some small changes to the NetCDF APIs, PnetCDF can use
MPI-IO to achieve high-performance parallel I/O.
More extensive installation guides can be found in file INSTALL. Additional
information regarding the contents of the release can be found in the
RELEASE_NOTES file in the top-level directory. Finally, the PnetCDF project web
sites below contain information on bug fixes and software releases.
https://parallel-netcdf.github.io
The previous project web page is now retired and has been migrated to Github.
http://trac.mcs.anl.gov/projects/parallel-netcdf
Requirements:
- PnetCDF relies on MPI-IO. ROMIO, the most common MPI-IO implementation,
can make use of a prefix in front of the file name to specify the
underlying file system and override ROMIO's auto detection. A typical
prefix would look like "nfs:" or "pvfs2:". Bear this in mind if you use
a file name with ':' characters. In some cases, ROMIO might think you
are passing in a prefix and will complain about an unsupported file
system type. If that happens, add a file-system prefix to your file
name.
- To build PnetCDF, you will need the following additional programs:
- autotools (autoconfig, automake, libtool, m4, etc.)
- Optional: (for PnetCDF developers)
- either yacc or bison
- either lex or flex
These are usually part of your operating system's development tools.
- To build PnetCDF with NetCDF-4 feature, you will need the following
additional libraries:
- NetCDF-4: https://github.com/Unidata/netcdf-c
- HDF5: https://portal.hdfgroup.org/display/support
Build recipes:
The process of building and installation of PnetCDF library is usually
simple: by just running commands "./configure", "make", and "make install".
(If more than one set of MPI compilers are available on your system, adding
option --with-mpi=/path/to/MPI/install to the configure command line is
recommended.) Several configure options are also available to customize it.
See INSTALL for more information.
Several machine- and compiler-specific build recipes are available under
the directory of doc.
- README.IBM (BlueGene systems)
- README.SX (NEC SX systems)
- README.CRAY (Cray systems)
- README.INTEL (Intel Compilers)
- README.LINUX (GNU compilers)
- README.Fujitsu (Fujitsu systems)
- README.K-Computer (the K computer @RIKEN)
- README.SGI (SGI systems, such as Endeavour @NASA)
- README.Ubuntu (Ubuntu)
PnetCDF User Documents:
C API References
http://cucis.ece.northwestern.edu/projects/PnetCDF/doc/pnetcdf-c
Questions & Answers
http://cucis.ece.northwestern.edu/projects/PnetCDF/faq.html
Note on supporting large files and large variables.
- Some Fortran APIs contain arguments of 64-bit integer data type, which
corresponds to MPI_Offset data type in their peer C APIs. Declare those
arguments as 'integer(kind=MPI_OFFSET_KIND)'.
- In places where you might use NF_UNLIMITED to define an unlimited
dimension in one of the nfmpi_ routines, use NFMPI_UNLIMITED.
NFMPI_UNLIMITED has been defined as the proper 64-bit integer type for
nfmpi_def_dim, whereas NF_UNLIMITED might be too small. Similarly, for
Fortran 90, use NF90MPI_UNLIMITED instead of NF90_UNLIMITED.
Mailing List
PnetCDF user community discusses the design and use of the library on the
[email protected] mailing list. Anyone interested in developing
or using PnetCDF is encouraged to join. Visit the list information page
https://lists.mcs.anl.gov/mailman/listinfo/parallel-netcdf
for subscription instructions. This mailing list is also for announcements,
bug reports, and questions about PnetCDF software.
The past discussions in the mailing list are available in the follow URL:
http://lists.mcs.anl.gov/pipermail/parallel-netcdf/.