When reading and writing case and data files in serial or
parallel, you can optionally use the Hierarchical Data Format (HDF). To read or write
case/data files using HDF, you can use the same GUI or TUI commands as you normally would
and simply append .h5 to the file name. Alternatively, you can
select HDF5 Case Files or HDF5 Data Files from the
Files of type drop-down list in the Select File
dialog box. If you want to write HDF files automatically at specified intervals during a
calculation, do not append .h5 to the file name in the Autosave Dialog Box, but instead set HDF as the default for reading and
writing files by using the following text command:
file → hdf-files?
Note: HDF files are not supported in Fluent Meshing or when running Fluent under ANSYS Workbench.
HDF files are always binary and make use of built-in compression. Thus, they cannot be viewed in a text editor. However, third-party tools are available that allow you to open and explore the contents of files saved in HDF format.
You can specify the I/O mode and/or the HDF compression level through text commands:
> file hdfio-options compression-level
compression-level [0-9]? [1] 4
> file hdfio-options io-mode
HDF5 I/O Modes:
1. HOST
2. NODE0
3. PARALLEL INDEPENDENT
4. PARALLEL COLLECTIVE
Enter Mode Number: [2] 1
HOST mode selected.
The compression level can be set from 0 (no compression; fastest) to 9 (maximum compression; slowest); the default level is 1. Note that increasing the compression level may not necessarily yield a corresponding decrease in file size.
The available I/O modes are:
| HOST: I/O is done serially by the host process. This is the default mode when the file path is not accessible to compute-node 0. |
| NODE0: I/O is done serially by the node 0 process. This is the default mode when the file path is accessible to compute-node 0. |
| PARALLEL INDEPENDENT: I/O is done in a parallel file system using the independent mode of MPI I/O. |
| PARALLEL COLLECTIVE: I/O is done in a parallel file system using the collective mode of MPI I/O. |
The Parallel I/O modes do not support compression. If you have specified a non-zero level for compression and you switch to one of the parallel modes, the compression will be set to 0.