This resolves #72 in an extremely simplistic fashion. The parts that I don't think will change is how the throttle count is passed from the config file to the solver/verifier.
For now it simply checks the count of tasks it is currently working with before commiting to solve/verify, and makes sure to delete those tasks after finalization. The next iteration would ideally differentiate between idle tasks and non idle tasks with the calls to the VM and RAM being the biggest bottlenecks for task throughput.
This resolves #72 in an extremely simplistic fashion. The parts that I don't think will change is how the throttle count is passed from the config file to the solver/verifier.
For now it simply checks the count of tasks it is currently working with before commiting to solve/verify, and makes sure to delete those tasks after finalization. The next iteration would ideally differentiate between idle tasks and non idle tasks with the calls to the VM and RAM being the biggest bottlenecks for task throughput.