First, to illustrate a point, I am going to assume a scenario with two separate physical disks, not virtual disks: In that case, in theory yes, in practice, I would say not so much any more with modern disks being so fast.
On Microsoft tests, the answer for questions on how to set up Exchange and SQL Server was always: Data on one physical drive (or RAID), log files on another physical drive. This made a lot of sense when drives still were transferring at 66 Mbps-133 Mbps and spun at 5400 RPM...or even much slower than that at one time.
Now with drives that are so fast (6Gbps and 15,000 RPM), while you may be able to measure the improvements with special equipment, I doubt you will notice great practical improvement by moving the swap file to another drive. We are probably talking about milliseconds here.
Now lets talk about two virtual disks: Probably no benefit at all if the virtual disks both reside on the same physical disk sub-system. The main benefit in the first scenario was to have two physical disks (or set of disk in RAID) work simultaneously on different parts of the same task. In this case, you are asking for the same one physical underlying disk (or RAID volume), which is often the case with a hypervisor, to do all the tasks, regardless of how your broke up the virtual disks.
swap
, notswop