How does CRuby manage the relationship between the interpreter, GC, and object heap as a unified runtime system?

CRuby integrates interpreter, GC, and heap through shared VM state, unified object representation (VALUE), and tightly coupled allocation paths.

CRuby is not modular between interpreter and memory system; instead, both are co-designed around a shared runtime state (rb_vm_t). The interpreter executes YARV bytecode while directly allocating objects into GC-managed heaps via fast allocation paths. Every object is a VALUE referencing RVALUE slots. GC can only operate safely when the interpreter reaches safe points, coordinated through VM flags and GVL control. This tight coupling ensures correctness but limits independent scaling of subsystems.