In practical applications, costs are often not amortized.

If lucky transactions are more than fast enough but unlucky transactions time out because they perform a rebuild it's a defect that cannot be accepted because average performance is good.

Most databases need to take care of worst case performance with techniques that increase complexity and degrade best-case and aggregate performance (e.g. maintaining two hash tables, with lookup accessing both, insertions into the preferred one, and background threads that move records from the old table to the new table).

I may have misspoken and should have left out "amortized," perhaps. It doesn't sound like you're familiar with incremental resizing or other real-time techniques?[1]

> Some hash table implementations, notably in real-time systems, cannot pay the price of enlarging the hash table all at once, because it may interrupt time-critical operations. If one cannot avoid dynamic resizing, a solution is to perform the resizing gradually.

[1]: https://en.wikipedia.org/wiki/Hash_table#Alternatives_to_all...

It's amortized O(1) even without any special incremental resizing. I think the goal of incremental resizing is to get closer to actual O(1).

But I'm not sure incremental resizing can get to actual O(1). Is the malloc() call for the new memory O(1)? And then after malloc() wouldn't the memory have to be initialized somehow (maybe just 0-initialized)?

malloc() isn't O(1), but allocations in a normal GC are O(1).

Is there some reason that malloc() would be slower asymptotically than a GC? Why couldn't whatever technique is used for the fast GC be used for malloc() as well?

GCs are typically allowed to (and often do) move objects, while malloc/free do not. This means that malloc/free must maintain data structures tracking allocated vs. free space, while compacting GCs keep all the free space in a single contiguous address space, so allocating an object just increments the "start of free memory" pointer.

Of course, that just moves the cost from the allocation side to the GC side. This means that the GC must maintain data structures tracking the type of each object (to know which fields are references), the root references, and several others (like card tables) for an efficient implementation. For malloc/free, releasing an object can just add it to a free list (a pair of pointer writes), while a compacting GC has to update all references pointing to the live objects it moves.

Yep! And this a pretty specific design element that I don't think gets appropriate attention when introducing GCs as a concept. There's a big difference between saying malloc vs. gc is doing "more work" or "less work" implying faster and slower.

Doing more work but elsewhere can (possibly) make your program faster. GC's often eat more CPU and RAM which hurts when you need more of either. But C/C++ can occasionally put malloc pressure on your critical path hurting you when you have plenty of resources where a GC might actually help you.

If it was any simpler it wouldn't be so damn interesting and fun.

malloc implementations often involve some kind of a search to find a free area that can satisfy the request. See http://www.gii.upv.es/tlsf/ for an O(1) implementation - use a large number of free lists, one per size, using bit scan instructions on the block size to select the list to use. To malloc, grab the first entry, and put any remaining portion on the appropriate free list.

To free, add the block to the appropriate free list. Assume blocks participate in a doubly-linked list of all blocks, used and free alike, in address order - adjacent free blocks can be coalesced at this point.