Optimizing your Linux server for memory-based NoSQL databases (Part I)

SLUB is already the default choice for an allocator and so is the sparse memory map.

It is worth pointing out that NUMA options won't help at all unless the database is carefully designed to operate on multiple cpus and balance the memory use among them equitably. Such problems are, sadly, non-trivial.

I don't see how any of these options will give any performance enhancements at all.