Initial implementation to add parallelization.

[kshoga1]

Initial implementation to add parallelization.

Credit to michaelmckinsey1 using GPT5 for the bulk of the work on this implementation.

[kshoga1]

Ran a couple performance tests between serial and parallel, and including the summary below. The dataset and testing are based on the what is used in the thicket tutorial notebook 08A_composing_parallel_sorting_data.ipynb (https://github.com/LLNL/thicket-tutorial/blob/develop/notebooks/08A_composing_parallel_sorting_data.ipynb)

Summary:

Test 1:

12,916 files

Serial:
Reading Thickets: 2min 23 sec
Creating Thickets: 2min 47sec
Total runtime: 5 min 14 seconds

Parallel:
Reading Thickets: 28 seconds
Creating Thickets: 53 seconds
Total runtime: 1 min 23 seconds
Improvement: 73.5% faster

Test 2:

25,832 files (test 1 file set duplicated)

Serial:
Reading Thickets: 4min 52 seconds
Creating Thickets: 5min 47 seconds
Total runtime: 10min 54 seconds

Parallel:
Reading Thickets: 1min 4 seconds
Creating Thickets: 2min 5 seconds
Total runtime: 3min 13 seconds
Improvement: 70.4% faster