RVI-Enabled CPU Increase Performance

If you are thinking of buying a new server for your virtual data enter you need to read this. VMware has release a whitepaper that shows that by using a server that has AMD processors with RVI-enabled CPUs you can get up to 40% in performance gains for MMU(memory management unit)-intensive benchmark and up to 500% for MMU-intensive micro-benchmark.

Just like in the 1990s when chip manufacturers start adding extensions in the chip sets to support multimedia so today Intel and AMD are beginning to add hardware extension that support virtualization. In 2006, both Intel and AMD introduced their first-generation hardware support for X86 Virtualization with AMD-Virtualization (AMD-V) and Intel VT-Technologies.

{loadposition txtbanner}

Using the recently release AMD second generation hardware that support MMU (Memory Management Unit) virtualization, called Rapid Virtualization Indexing (RVI) VMware was able to see the above mentioned performance gain.
Prior to the introduction of RVI, the VMware virtual machine monitor (VMM) used software techniques for virtualizing x86 processors. The VMM uses a method called shadow paging; “In shadow paging the VMM maintains PPN (physical page number) to MPN (Machine Page number) mappings in its internal data structures and stores LPN (logical page number) to MPN mappings in shadow page tables that are exposed to the hardware. The most recently used LPN to MPN translations are cached in the hardware TLB. The VMM keeps these shadow page tables synchronized to the guest page tables. This synchronization introduces virtualization overhead when the guest updates its page tables”. This overhead is negated when ESX VMM and VMkernel aggressively try to use large pages for their own memory when RVI is used.
*Performance of Rapid Virtualization Indexing whitepaper