Copy h5bench_write_normal_dist.c as baseline.

Created a new benchmark test name and added selection
support into h5bench.py test wrapper.

Added parsing of DATA_DIST_PATH into params struct to record
the data distrbution based on an input file.

Takes the provided DATA_DIST_PATH file and reads the data sizes
per rank from the file feeding into holder array.

Create make targets and add file to install list.

Remove rank 0 output summary limit so all processes report
performance rather than just accepting rank 0 results.  This
is especially important for non-uniform data distribution but
should be considered for uniform since no summary stats are computed.
Recommend --tag-output to track per-rank stdout.

Remove naive total size compute and limit it to a per rank size value.

Add scaling paramter to tests to scale particle count in order
to create memory footprints that more accurately reflect a data
distribution.

Data distribution inputs are in particle counts.
Particles are 32-byte structures so a data distribution measured in
bytes needs to be scaled down so the particals instiated match
the actual data footprint, in multiples of 32-byte particles.

Change the deployed binary to "write_var_data_dist" so it matches
the configured test reference in h5bench.
The binary name needs to match the test name so that it can be called
by the hbench wrapper.
This follows convention of "write_var_normal_dist".
Update code to log the correct benchmark name.

Change action/artifacts-upload from v2 to v4 to remove
dependence on deprecated v2.
The v4 syntax for the artifact-upload remains the same
so a simple update of the version number should be sufficient.

See blog post for details:

https://github.blog/changelog/2024-02-13-deprecation-notice-v1-and-v2-of-the-artifact-actions/

Update to latest version of container to see if it avoids the
missing distutils dependency reported with the @0.11 version.

…ama/h5bench into feat-data-dist

Grab the github actions fixes to ensure current feature branch validation tests
run successfully.