It's probably a lot more than that. A well-optimized bytecode interpreter is not 100x slower than native code. But also e.g. QuickJS uses refcounting with the o

It's probably a lot more than that. A well-optimized bytecode interpreter is not 100x slower than native code. But also e.g. QuickJS uses refcounting with the occasional tracing to remove cycles, and while it's a simple and reliable approach, it's not fast.