There is a self-assembled server with 4-core Xeon E3-1220 v6, 32Gb RAM. The challenge is to place the 2 Windows on each will be installed different software, which requires for normal work of 2 cores/8 gigs, but the full extent of the CPU's resources are utilized very rarely, if ever. Plus there Linux machine, which also requires 2 core/8 gigs, it will be more intensive CPU usage, but not extraordinary. And additionally, you need to place 3-5 light Linux new path, each under its task, but in any case they will a single core/1 gig. RAM's, like, no problem - ESXi allows you to cut how much you need until the end, but with the cores doesn't work - almost every dev to specify the use of 4 cores, but I don't quite understand - is it possible to do so and to push nevpihuemoe? The hypervisor somehow balancing, distributes the load, or there is another principle working?