The term "sockets" for a VM is exactly the same as a "socket" in a physical server, and the number of "cores" is per-socket, rather than total - indeed in later vSphere versions, this has been clarified in the VM settings UI:
A lot of in-depth discussion about cores vs sockets (and the effect on performance), is in this VMware blog post, specifically:
When creating a virtual machine, by default, vSphere will create as
many virtual sockets as you’ve requested vCPUs and the cores per
socket is equal to one. I think of this configuration as “wide” and
Assuming your example of VM with four vCPUs, your VM will see - by default - four physical single-core CPUs.
Because of licensing constraints (software may be licensed per CPU socket for example), you may want to keep the core count the same, but reduce the number of logical CPUs visible to the system - equivalent to your physical architecture.
When you must change the cores per socket though, commonly due to
licensing constraints, ensure you mirror physical server’s NUMA
topology. This is because when a virtual machine is no longer
configured by default as “wide” and “flat" ...
That said, the article goes on to explain that you may run into performance issues if you take this approach:
This configuration, which resulted in a non-optimal virtual NUMA
topology, incurred a 17% increase in execution time.
It's worth noting that these tests were run against AMD processors using NUMA technology, so results against Intel platforms may be different, especially if your CPUs do not have the NUMA capability (this was introduced with Nehalem in 2007) so YMMV.