Integrate PalsMpiexecSettings into Experiment factory methods

[MattToast]

Registers the PalsMpiexecSettings class in SmartsSim's Experiment factory methods so that the correct _BaseMpiSettings class is returned depending on the launcher the user is intending to use:

>>> pbs_exp = Experiment(launcher="pbs", ...)
>>> rs_1 = pbs_exp.create_run_settings(run_command="mpiexec", ...)
>>> type(rs_1)  # returns a OpenMPI compliant RunSettings
<smartsim.settings.mpiSettings.MpiexecSettings object at 0x7f9fe02eee00>
>>> pals_exp = Experiment(launcher="pals", ...)
>>> rs_2 = pals_exp.create_run_settings(run_command="mpiexec", ...)
>>> type(rs_2)  # returns a PALS compliant RunSettings
<smartsim.settings.palsSettings.PalsMpiexecSettings object at 0x7f9fdaf6f1c0>

[codecov]

Codecov Report

Merging #343 (c12c39f) into develop (f9e17f0) will increase coverage by 0.91%.
Report is 3 commits behind head on develop.
The diff coverage is 100.00%.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #343      +/-   ##
===========================================
+ Coverage    87.31%   88.23%   +0.91%     
===========================================
  Files           59       59              
  Lines         3531     3552      +21     
===========================================
+ Hits          3083     3134      +51     
+ Misses         448      418      -30     

Files Changed	Coverage Δ
smartsim/_core/control/controller.py	84.15% <ø> (ø)
smartsim/_core/launcher/launcher.py	100.00% <ø> (ø)
smartsim/database/orchestrator.py	83.91% <ø> (ø)
smartsim/settings/palsSettings.py	89.13% <ø> (+2.17%)	⬆️
smartsim/settings/settings.py	75.00% <100.00%> (+0.80%)	⬆️

... and 8 files with indirect coverage changes

[ashao]

Thanks for getting this in!

[al-rigazzi]

LGTM, but I gotta ask if we ran the tests on any PALS machine because I'm the baaad guy

[al-rigazzi]

Should we also support aprun here?

[al-rigazzi]

My suggestion is not yet, as I remember there were some issues. But we will have to in the future.

[MattToast]

I think I'm inclined to agree as I didn't specifically test against aprun. I'll be sure to throw in a ticket to track the future work needed!

[MattToast]

@al-rigazzi I manually scaled down a couple of the on_wlm tests to run on an internal 2 node EX/PBS/PALS machine, and everything seemed to be working as expected, but I was unable to run the test suite in full.

LMK if you think we should hold this off until we can do a more substantial test run!!