Currently, the allocators that are implemented will allocate their own heap allocation type separate from the concerns of the GcBox. Each of these allocations can potentially have their own header. This also means that we are potentially creating more overhead per allocation than needed where we have a header for the allocator box and a header for the GC box.
Ideally, we should reconcile these wherever possible.
Additional questions that may be relevant to address / think through:
- How does reconciling the allocation change with different GC policies.
Currently, the allocators that are implemented will allocate their own heap allocation type separate from the concerns of the
GcBox. Each of these allocations can potentially have their own header. This also means that we are potentially creating more overhead per allocation than needed where we have a header for the allocator box and a header for the GC box.Ideally, we should reconcile these wherever possible.
Additional questions that may be relevant to address / think through: