+/*
+TODO
+ Some tests are much more expensive than others (on cycles and synchronization), add a calibration step to figure out duration.
+ Some tests are more disrruptive than others, add weights to tests so some are picked more frequently.
+ The workload is too static, add some background generation noise tests.
+ The workload is too stable, add perturbance on the number of available tasks.
+ Fuse threadpool with non threadpool workloads by firing some tasks as separate threads.
+ Add an external watchdog so we can run the stress test in a loop for very long times and get results out of it.
+ We don't have enough tests, add one per locking operation we got in the runtime.
+
+Missing tests:
+ Ephemerons
+ Dynamic methods
+ Regular SRE
+ Remoting / Transparent Proxies
+ Context locals
+ Thread locals
+ Finalizers
+ Async socket IO
+*/