Optimize rax size tracking

[zuiderkwast]

Follow-up to #688.

Idea 1: Use the void *zmalloc_usable(size_t size, size_t *usable) which both allocate memory and return the allocation size, so it can be used in a single call instead of zmalloc and later calling 'zmalloc_size'.

Idea 2: Use the diff of the per-thread used_memory_thread in zmalloc, to see the memory usage before and after each rax operation and calculate rax used memory delta from this. This only works for up to some max number of threads, if we don't change this in zmalloc.c, but there is no clear answer to what to do if some module creates a lot of threads.

[zuiderkwast]

@ranshid @knggk FYI