So it sounds like it is not so much the allocation/de-allocation savings that matters, but rather that you get to skip initializing the memory because it is already known to be a valid instance of a given struct type?
See also SLAB allocation, used in many OS kernels: https://en.wikipedia.org/wiki/Slab_allocation
So it sounds like it is not so much the allocation/de-allocation savings that matters, but rather that you get to skip initializing the memory because it is already known to be a valid instance of a given struct type?