h5dwalk

The h5dwalk utility provides a parallel alternative to creating and running script based approaches to invoke serial HDF5 tools on a collection of hdf5 files.   As a means of invoking parallel instances of a serial tool, the h5dwalk application can accept directories as input arguments.   This new tool provides recursive file discovery and filtering to select hdf5 formatted files.   The resulting file collection is distributed between MPI ranks and individual files are then selected for input to a user selected application.  Below are the command-line options for h5dwalk. For more information we refer you to User's Guide.

[bin]$ ./h5dwalk --help

Usage: ./h5dwalk [options] <path> ...

Options:

  -i, --input <file>      - read list from file

  -o, --output <file>     - write output summary to the named file.

  -E, --error  <file>     - write processed errors to file in text format

  -l, --log_text <dir>    - write individual tool outputs to a file. 
                            Logs can be written to an optional named directory.

  -T, --tool <executable> - name of the HDF5 tool to invoke

  -h, --help              - print usage

For more information see https://mpifileutils.readthedocs.io.

The example shown below is run using 4 cores and generates a log file with the name “showme-demo.log”.  

[ bin]$ mpiexec -n 4 ./h5dwalk -i ../../tools/test/demo-h5dwalk.txt -o showme-demo.log

The tool was added in 1.13.0 release.