A baseline comparison of the same tests running in Boot Camp might give a nice understanding of what kind of general performance loss you're going to get by virtualising at all.