[FEATURE REQUEST] Use multi-threading to speed up benchmark plotting about gcpy HOT 2 CLOSED

I put together something functional using the joblib package a while back that executed plot creation for one collection per thread. It was over a year ago so I used gcpy prototype code in a Jupyter notebook that I had not yet put into the gcpy repo. The non-parallelization part is all very outdated at this point, and I did some extraneous stuff such as storing some input info in a config file. But it did work for parallelization and I was able to create a full suite of collection comparisons within minutes. I am not sure if joblib is the best option, then or now, but here are the files for reference.

Job submit script:
make_benchmark_plots.run.txt

Script called in job:
plot_collections.py.txt

Config file:
config_bench.py.txt

Script with functions used (can now just use gcpy):
v1102e_1hr.py.txt

from gcpy.

Parallel functionality has been added to compare_single_level and compare_zonal_mean, where variables are now plotted and merged into a single PDF in parallel. Parallel functionality has also been added to several benchmark functions, where separate PDFs for different categories of variables are created and populated in parallel. These changes have resulted in a 2-4x speedup of benchmark plotting.

from gcpy.