Parallelization at backend board level

Summary

Moved here from GitHub issue:

Current loops that operate on hardware during state transitions loop over all hardware devices. This should be adapted to parallelize operations done through a single back-end card (AMC), and only perform loops where parallelization is not actually possible/beneficial (e.g., OptoHybrid/VFAT)

Current Behavior

Loop is made over AMC objects, actions are performed sequentially

What is the expected correct behavior?

Perform per backend board object actions in parallel