Using large pages in virtual machines

VMware ESXi provides 2MB memory pages, commonly referred to as "large pages" along with usual 4KB memory pages. ESXi will always try to allocate 2M pages for main memory and only on failure try for a 4K or small page. Virtual machines are large pages if 2M sequences of contiguous Memory Page Numbers are available. The idea is to reduce amount of page sharing and also increase the memory footprint of the virtual machines. The biggest benefit is of mitigating TLB-miss (Translation Lookaside Buffer) costs as much as possible for Nested Page Table enabled servers running ESX.

However, allocating memory in 2M chunk may cause the memory allocated to the VM to become fragmented. But as small pages are allocated by a guest and VM, these larger sequences need to be broken up.

So, if defragmentation occurs there could be enough memory to satisfy a large page request even when there is no 2M contiguous Memory Page Number's available. The defragmenter's job is to...