What is the GIL switching interval? How does the CPython interpreter decide when to acquire the GIL lock from one thread and pass it to another, if both threads are performing pure CPU-bound computations?