1. Which of the following scheduling policies is preferred for a large VM, which is configured with more virtual processors than the physical cores on a single NUMA node?

A. Schedules all virtual processors randomly, to have an even distribution.

B. Schedules all virtual processors to a single NUMA node so that the guest OS can treat them uniformly.

C. Splits the virtual processors into groups (virtual NUMA nodes) that are placed to separate physical NUMA nodes, maps VM memory interleaved among the NUMA nodes, so that the guest OS does not have to be NUMA aware.

D. Splits the virtual processors into groups (virtual NUMA nodes) that are placed to separate physical NUMA nodes, exposes the virtual NUMA topology to the guest OS and have it handle the memory allocation.