Comment by userbinator

It's great to be able to make memory allocators faster, but it's also wise to consider the possibility that it's the applications that use them which are where the inefficiency is - and ironically it's the applications that do the most unnecessary allocation/deallocation that would benefit the most from a faster memory allocator.

To paraphrase a common adage, the most efficient way to allocate memory is to not do it at all. Probably helps with reducing the chance of use/free bugs too.