What is the point of a 64-bit chip only in ML?

jbg232 said:

I'm really just wondering here, given that 32-bit applications are here for a long time, what really is the point of limiting the ML OS to 64bit chips as there will still have to be backwards compatibility for a long time. Will it really make the machine faster? I understand the exclusion of GMA950 and X3100 integrated graphics GPUs as ML will likely support much higher graphics resolutions in the future, but the 64bit thing I just don't get. Please educate me.

Click to expand...

It'll make the OS and the rest of the apps a bit faster by having everybody focused on optimizing purely for the 64-bit operations.

The idea is that they no longer have to allocate any of their engineering resources to optimize OS X for 32-bit chips as well. You can't simply recompile the OS from 32-bit to 64-bit operations or vis versa and expect it to be just as good.

Also, 64-bit CPUs have additional features that 32-bit CPUs can't use, now they can fully start to take full advantage of all 64-bit instruction sets without having to worry about the impact on the 32-bit CPUs.

I would guess that it is a way to kick developers (especially driver writers) into coming into the present. Other than that, the only benefit I can see is that Apple gets to stop compiling a 32-bit kernel, which doesn't seem like that much of an overhead to me.

Of course, it could also be a way for Apple to sell more Macs, if people on old Core Duos want to upgrade to ML.

With Windows, it makes more sense, since 32-bit Win kernels can't run 64-bit apps. With Win 2008 Server R2, and Exchange 2007, it is 64-bit only.

32-bit OS X kernels CAN run 64-bit apps, so either it is money-related, development-related, or Apple is just trying to get everyone to the present.