elevate said:
And unlike another company's so-called 64-bit OS, this one actually is 64-bit.
And you think that's an advantage? Time for a quick lesson in computer architecture....
First, Mac OS X isn't the only OS that does this. Solaris can run a 32-bit kernel on 64-bit hardware. So can Linux. This is actually fairly common usage among sysadmins, as 64-bit kernels tend to be inherently slower. If a 32-bit kernel has a 64-bit VM, there's no real advantage to running a 64-bit kernel unless you have drivers that somehow benefit from a huge address space. (I can only think of two---Infiniband and one other similar type of device that I can't remember.
)
The 64-bit AMD architecture is a -very- unusual architecture. On the x86-64 architecture, 64-bit code runs faster than 32-bit code. The -only- reason that this is the case is that the IA32 (x86) architecture is -severly- register-starved to begin with. Running the chip in 64-bit mode causes it to have access to twice as many general-purpose registers, so it isn't (as) register-starved (16 GPRs instead of 8).
On most architectures, however, 64-bit code is slightly slower than 32-bit code unless it is doing significant amounts of 64-bit integer math (which is not common in an OS kernel) or is able to take advantage of the extra address space in some useful way (e.g. working with very a large database in RAM).
Apple made the only logical choice, given their hardware (32 GPRs on PowerPC), and Microsoft made the only logical choice, given their hardware (8 GPRs on x86-32, 16 GPRs on x86-64). They made different choices because the underlying hardware has fundamentally different performance characteristics. Both were, in my professional opinion, the right choices.