Re: Garbage Collection in Ada

[dewar@merv.cs.nyu.edu (Robert Dewar)]

1.  The worst-case memory requirements of ANY general purpose malloc
implementation are quite bad.  This is a mathematical theorem.  The only
way to limit this is to restrict object sizes.

   Yes, and restricting object sizes is a common technique. Indeed
   going all the way and restricting object sizes to a single size
   completely eliminates fragmentation, but does NOT eliminate the
   problem of CGC hanging on to blocks that should be freed.

3.  Vendors rarely specify the allocation algorithm.  So it's not
possible to tune for the algorithm even in nonportable code.

   Yes, and providing your own allocator with known characteristics
   is a common technique.

- "Virtual origins" work fine with a conservative GC, so long as you
also keep the naive pointer representation.  This is likely to be cheap
(at most one extra store).

   Actually it is a space issue more than a speed issue, it means that
   pointers to unconstrained arrays would need to be three words instead
   of two, certainly not fatal, but not space that would seem worth
   spending for the relatively specialized use of 3rd party CGC's.