Simulation Threading

[JJL772]

This is supposed to work around timeout issues seen when running Badger against Linac-Simulation-Server.

My understanding of the issue is:

• Writes to any PV trigger a full re-evaluation of the model (on my system nc injector takes, ~6s)
• While the model is being re-evaluated, the server becomes unresponsive
  • This leads to caput reporting a timeout when it tries to provide a readback of the new value
  • The epics interface in Badger may timeout because it only waits ~2 seconds while attempting to get a readback value
    • Additionally, the constant re-evaluating of the model on every single write will slow the optimization to a crawl

There are two ways to fix this:

1. Go back to the previous implementation that would evaluate only on read (this would break PVA support and CA monitors)
2. Implement some type of threading to prevent the server from becoming unresponsive.

I opted for option 2 just to avoid breaking PVA.

VIRT:BEAM:SIMULATE_TIMEOUT defines the window in which PVs can be batched before the simulation is re-run

VIRT:BEAM:SIMULATE will force the simulation to occur regardless of VIRT:BEAM:SIMULATE_TIMEOUT setting. This is mostly handy for working with the server from the command line.

With the timeout method, no changes are required to Badger-Resources

[roussel-ryan]

Please compare this to the calculations done in lcls-tools screen.image to make sure it is consistent. I believe it is reversed