Merging NREL updates

[kevinmenear]

I’ve prepared a snapshot import of the NREL fork into this public repo on branch migration/private-commits. It’s been a while since this code was tested on other datasets/clusters. This issue is a central place to document the major updates since the fork, coordinate testing, discuss gaps, and capture required changes before we merge into master.

Next steps

• Document major updates
• Collect feedback and test results here.
• Address action items
• When we’re confident, merge migration/private-commits to master

[kevinmenear]

Suggestions for initial test:

1. Pull the updates on the migration/private-commits branch
2. Update the default_config.yaml file based on the system and data being simulated. You will need to update considered_partitions, sim_start, sim_end, and Pdefault. You can make impromptu_reservation_names an empty list, though it should work even with those dummy names in there. The system_name can stay "default", that will process node names based on the original implementation (before the fork).
3. resv_dump_historic is new. It pulls in historic reservations from slurm, beyond the reservations in place when the slurm_dump.sh script is run. This is a result of the new slurm_dump.sh script. If that data is not available, you can just make a blank csv at the filepath pointed to by that parameter, with the columns Id|Name|Nodes|Start|End|TotalTime|
4. The various other parameters may need to be tweaked based on the system, but that should only influence accuracy, and the simulator should run to completion regardless of what those parameters are set to.