Explore adopting sys.monitoring (from Python 3.12, PEP 669) to accelerate memory profiling

[emeryberger]

Is your feature request related to a problem? Please describe.
Scalene currently uses tracing as part of its memory profiling logic via sys.settrace, which is a source of overhead.

Describe the solution you'd like
Explore an implementation using PEP 669, specifically sys.monitoring, which in principle should enable a lower overhead implementation of the same logic.