Apple and Intel sitting in a tree, K-I-S-S-I-N-G.
Well, sort of.
Does it seem likely?
Well, let’s put some facts we do know about on the table and see how they fit into the rumor/announcement made today.
Apple has been frustrated, to say the least, with the recent performance of PowerPC processors.
This has been well known for some time, and it’s been no secret that this has often been a trigger for rumours about Apple and Intel.
X86 CPUs seem to have been winning the speed race. Or more correctly the Mhz race. We all know that Mhz (or Ghz as we are now) is probably the least efficient and realistic way to measure CPU performance.
Even if an x86 version of Mac OS X was made available there would be no guarantee that it would be faster than the PowerPC equivalent. In fact, if I was in Apple shoes I’d be looking at the AMD 64-bit solutions (primarily the Opteron, which already exists in a dual-core format and used in multi-way machines like those from Sun. The Intel multi-core units, in comparison to the AMD multi-cores have not faired as well as Intel had been hoping.
Also, it seems unlikely that in the face of the announcements from Sony and Microsoft that they would be using IBM sourced chips, and PowerPC chips no less in the Xbox 360, that Apple would simply dump PowerPC.
Apple also use the HyperTransport bus, a technology associated with AMD, not Intel, for the G5. Using an Intel CPU may preclude Apple from using HyperTransport, a technology they are obviously happy with, and that would lead to longer and more expensive development of the new hardware.
Apple is no stranger to CPU architecture changes – the move from 680×0 chips to PowerPC was comparatively seemless, but that doesn’t mean that the same would be true with PowerPC to Intel.
Darwin is available for Intel/x86.
Getting the core part of your operating system supported on a new architecture takes you a long way towards supporting the entire OS. Darwin has always been available for x86, but that isn’t all that surprising since Darwin is based on an OS, which was almost entirely developed on x86 platform.
Apple also have much of the other parts of code supported under x86. There was a time when having an Apple application on a Windows platform was a rarity. Now we have iTunes, QuickTime and even smaller apps like the iDisk connector. As Windows focused applications these aren’t bad, and it wouldn’t take as much work as we might thing to migrate the opt layer of OS X – that is, the video components and Window interface that is Quartz and Cocoa in combination with the other tools that make up the layer that sits on top of the Darwin core.
Even if Apple ported the rest of OS X to Intel – and let’s ignore the fact that it might be for the purposes of a faster processor and is, instead, for compatibility reasons with Windows applications, there’s still more to do.
The underlying CPU is only part of the compatibility problem if Apple wanted to let users run Windows applications on Apple hardware. There are the libraries and other components that make up Windows in the same way that Quartz makes up OS X. Knowing Apple, they could be hoping to resolve this with the Apple equivalent of WINE, or even WINE itself, since Apple aren’t new to adapting free software to use in their systems.
CPUs aren’t the only thing Intel makes
Despite what you might read in the press, Intel make a heck of a lot more than simply CPUs. They also make controller chips for most types of hardware, including RAID, networking, and the core motherboard chips that make up the backbone of many of the Intel based hardware out there.
Apple also don’t use PowerPC chips in all of their equipment. The AirPort base stations for example use AMD chips, and there are plenty of other components from Apple (AirPort Express, iPod, to name a couple) that also don’t use PowerPC CPUs. I don’t know about the Apple RAID solutions, but Intel have a lot of experience with RAID, and especially using SATA drives.
Apple could be talking to Intel about developing a PowerPC compatible version of their Centrino technology, which combines networking, interfacing and most importantly wireless networking technology together into a single chipset. As a single chipset, it’s smaller, takes up less space and allows for smaller and lighter laptops.
Centrino also has advanced power management techniques that can help to eke out the absolute most from your battery power.
It’s feasible Apple could be looking to use Centrino like technology in it’s forthcoming PowerBooks and the combination of small size and enhanced battery life would make Centrino an ideal companion to a low power G5, which is going to need all the space (for heat dissipation) and power (for better than 30 minute) battery life enhancements it can get.
We have to temper every rumour that comes out of Apple, or about Apple, with a massive dose of salt. We’ve been wrong, or bitten, too many times to take every little factoid at face value without any more information.
Apple’s meeting with Intel could therefore be as innocent as looking for a new partner for some of their existing, non PowerMac/iMac/iBook/PowerBook technologies like Xserve RAIDs or AirPort hardware. This is where my money would lie.
Equally, they could be discussing Centrino for PowerPC as the next step forward. I think there’s a strong possibility here, because the G5 PowerBook is now a sore point in Apple’s side; their top of the line laptop is not so top of the line anymore.
But the least probable option has to be a move to Intel-based CPUs. Apple have too much R&D invested in PowerPC, in the PowerPC consortium and in non-Intel compatible technologies like HyperTransport for them to opt for Intel as a first choice.
If the company had been AMD, I think I might have come to a completely different conclusion.