That means number of GC cycles in this situation is proportional to number of objects you allocate, while each GC cycle time is proportional to the managed heap size. To avoid using the GC in Higgs, I'd basically have to implement my own specialized memory managers (my own allocation and garbage collection) for each of the data structures used in the compiler. Saying that I shouldn't use the GC is basically denying any responsibility the D GC might have and shifting the burden onto me, the application developer, to implement my own specialized memory management for the different use cases where I need dynamic memory allocation, and there are many such use cases.