Init/disposing of RhinoCore instead of start/stopping child processes

Hi, I’m trying to see if I can improve the response time of Rhino Compute when the child processes are not running.
I’d like to know if there was a reason for not running the child processes at the start, but without loading RhinoCore until the first request. Would this incur core-hour billing while RhinoCore is not loaded (but after the Resolver initialization)?
Same for the shutdown, would it stop billing as soon as RhinoCore is disposed, without terminating the child process?