So Apple’s cards have been on the table for a long time now. Mobile gear and traditional personal computers are two different breeds, and they are not meant to become one. At least not now. This is very different from what Microsoft is trying to do with tablets, witness the Surface, which has, so far, been a pitiful failure. With the Surface, you have two near-identical form factors, the RT version and the Pro version. The former uses the same ARM processors found in tablets, whereas the Pro version sticks with Intel Inside.
The problem arises when you try to use traditional Windows software on the RT brand. You can’t, and that’s been a source of endless confusion for customers who see two operating systems meant to look the same, but which are utterly incompatible. Microsoft’s focus has been on making PC note-books into convertibles. They have touchscreens that you use your fingers or a stylus to manipulate, along with the traditional keyboard and trackpad. Microsoft doesn’t consider that a normal note-book orientation can be a fatiguing exercise if you choose to use the touchscreen.
Certainly, many ads for the Surface tablets reveal a traditional note-book motif, as a result of the physical keyboard cover and the kickstand.
Apple’s approach focuses on two different form factors that aren’t meant to become hybrids, that you can’t mix a toaster oven with a refrigerator, and thus the operating systems must be different. Although iOS-inspired apps and a handful of interface conventions have shown up in recent versions of OS X, it’s still a Mac through and through. This is Apple’s game plan, although the day may come when the traditional personal computer is no longer needed by most people.
But what about the processor?
Apple has already successfully navigated through two major processor transitions without causing a whole lot of pain for customers and developers. The mid-1990s switchover from 68K to PowerPC did mean that developers had to rebuild apps, but Apple delivered an emulation layer to keep most of your older apps running, although performance was a little tepid till PowerPC gear became fast enough. The 2006 switchover to Intel also required emulation, via an app known as Rosetta. Once again, developers had to rebuild their stuff to be compatible with Intel, and some apps never made the transition. When Rosetta was vanquished beginning with OS X Lion, Mac users either had to stick with its predecessor, Snow Leopard, or get their apps updated.
So far, Intel has worked great on Macs, but Apple is into building custom chips for iPhones and iPads, based on ARM reference designs. Beginning with the A7, featured on the iPhone 5s, iPad Air, and the forthcoming second generation iPad mini, Apple boasted a 64-bit desktop class processor. But how close does it come to a traditional Intel Haswell chip? It has been suggested that the A7 is equivalent to a Mac of several years ago. Not bad, but putting one inside, say, a MacBook Air, would not yield a very snappy user experience, but it would surely boast great battery life.
But that’s today. The A7 is roughly twice as fast as the A6, and it all happened in a single year. Compare that to the current Intel chips, which are only a tad faster than their 2012 predecessors. The real improvements come in battery life and graphics. Intel has finally delivered an integrated graphics solution that comes close to that of some midrange discrete chips.
However, with the larger emphasis on power efficiency, without major performance improvements on the latest Intel processor refreshes, is it possible for Apple to build an ARM-derived chip that would match a traditional midrange Intel mobile chip? I don’t want to call this a cop-out, but I’m not going to go that far, except to suggest that it could possibly happen in the next few years. At what point does Apple consider an ARM switch, so the same processor family can be used across the board? Such a move would give Apple even greater control over the Mac’s destiny with custom-built chips.
Unfortunately, the tech pundits arguing in favor of such a move seem to forget the developer dilemma. Many Mac apps aren’t fully compliant with Lion, particularly the full-screen and auto save features. Now imagine having to go one step further, rebuild an app to talk to ARM. It may not be so hard if Apple makes the conversion tasks easy in Xcode. But just clicking a button is a fraction of the job. There are all sorts of code optimizations that will be required, and what about “fat” apps that work on both ARM and Intel?
And what about apps that are still Intel savvy? Does Apple build a new Rosetta-style translation utility, or do they embed some sort of chip-based conversation feature, to reduce the performance loss as much as possible? Now that Apple has control of chip development, I suppose most anything is possible.
So do I think it’ll happen? Well, even if Apple went all or mostly ARM on Macs, it wouldn’t necessarily result in lost business for Intel. According to published reports, Intel is going to start building ARM chips, so they could still keep Apple’s business if the need arises. Apple clearly wants to ditch Samsung as much as possible, so another assembly partner would be a good thing.
Print This Article