Nice!
I'd be interested to see if you actually see a performance improvement, however, from setting the number of cores lower than the maximum number (at least when you're using the host machine at the same time)... Reason being that in most hypervisors (I could be wrong here, so someone please correct me if I am...) the VM will only be scheduled CPU time when all cores are available concurrently.
If you're using the host machine at the same time, this is going to mean that whenever you're running even a single process with a single thread, the entire VM is off-CPU because in order for the guest OS to do that multithreading, it needs its full contingent of cores simultaneously (having no real awareness of the host system).
Again, my previous disclaimer about possible factual inaccuracy stands (I don't know what work has been done on hypervisors here within the last few years, and I wouldn't be surprised if VMWare player runs completely in userspace and therefore does some crazy weirdness), but it would be nice to see a comparative load test if you're set up for it!
All I know that on my 4690K I get exactly the hashrate I should be getting in the VM (compared to the same generation chip running only Linux) while everything else is idle.
Ya I noticed very little slowdown speed wise.
Cool! I wonder if VMWare player is doing anything particularly clever or if the Windows scheduler is just doing a really good job... Probably the latter to be fair.