What About An A6 Mac?
September 7th, 2011With Lion signalling an apparent merger of at least some iOS features on the Mac, there’s more and more speculation as to Apple’s end game. Do they ultimately intend to combine the two operating systems, thus offering a seamless experience among iPhones, iPads, and Macs? If that happens, would Apple, in turn, ditch Intel processors and replace them with Apple’s own processors based on ARM technology.
I suppose this seems sensible at first blush. After all, Apple’s mobile gadgets are incredibly fast, with near-instant boot and app launch times. So why not put the very same chips on, say, a MacBook Air?
Now when it comes to Apple, you certainly can’t predict what they plan to do six months from now, let alone two or three years. So nothing ought to be off the table, but some of this speculation doesn’t quite pass the logic test.
First and foremost, the reason Apple’s mobile processors seem so fast is that they aren’t being asked to do near the amount of tasks they’d have to perform on a traditional personal computer. The iOS and all those apps are optimized to operate with extreme efficiency on a computing device that runs at roughly the same speed as a Mac of, say, eight years ago. While it’s a sure thing Apple will be speeding up their A-series chips in the years to come, expecting them to match, or closely match, Intel’s processors seems a stretch. That would require Intel virtually standing still while the mobile processors close the gap.
Besides, why would Apple want to cripple Macs in that fashion? One of the big selling points of a Mac these days is virtually instant response for most any task you want to perform, except for heavy-duty rendering tasks that require the fastest multicore processors, and the speediest drives available. Certainly the growth of solid state drives on Macs would make up some of the performance gap, but not for processor intensive work. Besides, why would Apple force developers to undergo a new processor transformation without an upside?
What the pundits who talk of a Mac with an A6 or A7 processor (assuming that model designation will continue) seem to forget is that, while Apple has managed to make their operating system work on different processors with great efficiency and reliability, what about the software? An app that supports Intel would have to be recompiled to run on the ARM-family processors that Apple uses as the basis for their home-brewed chip designs. There are still apps out there that won’t even run on Intel processors, and that transition occurred in 2006. Apple has discontinued the Rosetta emulation app that allows you to run PowerPC software for Lion, and that’s one reason some can’t upgrade unless they can find replacement apps, or the apps they’re using are upgraded.
Sure, Apple can devise a new emulation layer to deal with such issues for the next processor migration, but, again, why? Sure, if Intel’s processor roadmap goes off the cliff in the next few years, Apple would seek alternatives. But they could consider AMD, who makes chips that are compatible with Intel’s in terms of running the very same operating systems. Yes, Apple’s motives may at times be inscrutable to outsiders, but what they do ultimately tends to make a lot of sense when you begin to see the goal post.
Granted Apple might want to do more to unify OS X and the iOS in the years to come, largely to make it easier for customers to move from one to the other. It’s also quite possible more and more Mac and Windows users will decide that the iPad suits them just fine and abandon the traditional personal computer. In fact, the main justifications for Apple’s alleged OS meld and ditching Intel processors is to accommodate the reality that fewer and fewer people will need a real PC. Traditional computers will be confined to so-called “prosumers” and business customers, and those numbers will probably decline in time.
But so long as Apple can make healthy profits selling several million Macs every quarter, why should they kill that market? Sure, maybe the future A6 and A7 will have huge advantages in power utilization, but it’s also clear Intel isn’t going to stop making their chips more and more power efficient. I suppose a lot of possibilities are on the table, but that doesn’t mean you can predict Apple is going to go into any particular direction. This is one company that has a habit of confounding the experts time and time again.
So it all boils down to this: If Apple can see a strategic reason to merge operating systems and to switch processors, it will happen some day. But Apple isn’t apt to want to force developers to switch gears all over again in the very near future.
Now when Lion’s successor arrives, perhaps two years from now, maybe things will change. iPad sales and how the Mac is doing at the time may be determining factors. But, as always with Apple, prepare to be surprised and maybe amazed.
Print This Post
I speculate that much of the interest in Apple switching from Intel to ARM processors in the Mac has more to do with geek politics, than any actual benefit to the user. Some folks like the idea of Apple providing an integrated Apple-only solution, while others think that Mac OS X and iOS will be fully integrated. There also may be some lingering distrust of Intel from the old RISC vs. CISC debate among veteran Apple users, or some still see Intel as part of “Wintel”, even though Microsoft and Intel have now largely divorced from each other.
I don’t think moving the Mac over to the ARM architecture makes much sense, at this time. The software transition would be difficult for Intel-only apps. When Apple moved from the PowerPC, they had a ready-made emulator that they could license from Transitive (now part of IBM), which became Rosetta. Apple would have to develop their own emulator, which is no simple task, and would also be a legal minefield with Intel. Part of Intel’s settlement with Nvidia included preventing Nvidia from creating an x86 emulator; Intel is very protective of its instruction set.
From a hardware standpoint, Intel makes the fastest mass-market processors available. There’s an argument that ARM could eventually scale to meet Intel’s performance, but keep in mind that much of ARM’s power savings would be negated as it began to increase in performance. Besides, many other components in a typical computer use more power than the CPU (such as the display and storage devices).
Lastly, you need to evaluate the Intel-Apple relationship as it currently stands. Apple is Intel’s “premiere” customer. Apple is the only computer maker that isn’t required to be part of the “Intel Inside” co-branding campaign, receives early access to Intel CPUs and chipsets before Intel’s other customers, and has worked with Intel on a number of new technologies (such as Thunderbolt and adopting EFI). Keep in mind that Intel doesn’t just provide a CPU, but has vast manufacturing capabilities, process technology, and infrastructure validation that Apple isn’t capable of providing. Intel has its flaws, but it is a reliable, predictable partner; something Apple didn’t have with previous CPU suppliers.
While it’s possible that Apple could, at some point, release a device that has similar functionality to a Mac with an ARM-based design, I find it highly unlikely that they will switch from the x86 architecture. The Mac is a stable, growing platform, with a mature operating system and toolset, that continues to receive increasing interest from developers (not to mention the substantial benefits of x86 virtualization software capabilities that don’t exist with ARM). Apple’s relationship with Intel is very good, with early access to technology. While it’s fun to speculate about such things, another architecture switch, with no obvious benefits, and many non-trivial downsides, doesn’t make good business sense. Also, keep in mind that Apple pays a royalty to ARM for each of their processors sold, as well as their CPU fabrication partner, so it’s not like they would be cutting out a third-party in changing the architecture.
In my opinion, as long as the Mac is Apple’s performance computing platform, they’ll stick with Intel.
Colstan
I think A-series processors are going to take over more and more of the low end Mac world as they grow more powerful, but that Intel will be working on faster, lower power processors for Macs. Apple is pushing them hard, and Intel is responding as demonstrated by their Ultrabook challenge. Our last big purchase decision was between an 11″ MacBook Air and an iPad. (We went for the MBA. We really like having a keyboard and have a few home brew applications we use, though the Clang/LLVM technology that Apple is using will eventually allow developers to produce run anywhere code.)
Apple seems to be doing to its user interfaces what Airbus did for its popular 300-series aircraft. A pilot sitting in the cockpit of an A320 or an A380 uses the same muscle memory to find controls which work more or less the same way and are in similar locations. The various displays use the same conventions and iconography and are positioned more or less just where they would be in any A-300 plane. In fact, the aircraft handle and fly quite differently. An A340 or A380 titan does not respond the same way a much lighter A310 or A320 does, and each aircraft has myriad idiosyncrasies in how it handles. A professional pilot needs to be qualified for each plane separately, but it is an easier process to move from an A320 to an A380 than from a Boeing 737-300 to a Boeing 747.
@Kaleberg, For Apple to use an ARM-based A-series processor, they’d have to devise emulation for Intel-based software, or some sort of chip-based emulation capability. Even then, apps that aren’t in their native environment would run slower. The other alternative would require support for the X86 instruction set in an ARM or A-series processor to manage this situation effectively. Otherwise, developers would be forced to build apps for both ARM and Intel, the whole “Universal” mess all over again.
This sort of schizophrenic hardware scheme doesn’t pass the logic test.
Peace,
Gene
@Kaleberg, In addition to what Gene stated above, Apple would be needlessly fragmenting their product line. Mac OS X and iOS are separate products, no matter how much “Back to the Mac” marketing Apple talks about. They’re pushing their ARM series for the iOS devices. The A-Series, specifically the iPad, are not taking substantial Mac sales; they are devouring the PC netbook category, much as the Mac is squeezing the PC market at the high-end.
Keep in mind that Intel is a component provider for workhorse computers. ARM processors are nowhere in that league. Even Microsoft realizes this. Windows 8 will feature an ARM port, but it will only be available to tablet partners according to the reliable Wall Street Journal; you won’t be able to purchase it on a standard PC, because it will lack the ability to run Intel software, as well as a lack of device drivers.
I’ve yet to hear a compelling argument for Mac-on-ARM. It’s a lot of vague speculation that never addresses the substantial hurdles involved in switching from Intel. Issues such as a lack of an Intel emulator, x86 virtualization support, a high-speed processor architecture, versatile chipset support, scaling into the high-end of the market, product fragmentation, and not to mention losing the ability to run Windows through Boot Camp (as I said, Windows 8 on ARM will only be sold through select OEM partners for tablet designs).
ARM-on-Mac is something that is amusing to speculate about, but when you have to deal with the actual technical and business aspects of a switch, the idea falls apart.
Colstan
The only ARM based “Mac” which makes sense to me at the present time is something built on the Macbook Air chassis. The great weakness of the iPad is that it lacks a real keyboard. I have seen any number of people hauling around various and sundry keyboards with their iPads. I have even seen a device that the iPad slips into which has a physical keyboard and looks an awful lot like a clamshell Macbook.
The Wintel word had a laptop some years ago which had a swiveling screen so that it could be folded flat and used as a touch pad or flipped around and used conventionally with the keyboard. The problem with it was that the OS poorly supported the touch screen function. I have seen recent Wintel products which have a screen that rotates within a frame to become a touch pad and back to become a conventional, but small, laptop.
If Apple were to put the iPad “guts” into a Macbook Air chassis and add some variant of the rotating top or screen to it I believe they would be surprised at the demand for the product. It would have the same or better battery life as the iPad as there is a little more space for a battery and could be used for honest to goodness document generation while still being compact, easy to carry and light weight. It would not put an end to the iPad, but would extend the capabilities of the iPad for those who need “more”, but not the full feature set of the Macbook Air.
By the time this could be brought to market, the A6 processor should be available.
@Richard, But also remember that all those Wintel tablets were utter failures, except in a few vertical markets. Our family doctor has some of them and they truly slow down office efficiency. But they are so embedded into the practice’s management system that change is difficult.
Peace,
Gene
@Gene Steinberg, Yes, Gene. That was not the point, however. That is not relevant to my comment as the product I suggest would be little more than an existing iPad stuffed into the form factor of a Macbook Air with, perhaps, a little more battery capacity. More battery capacity would be nice, though not absolutely essential.
The failure of the Wintel convertible tablets was not the cause of the idea, but the implementation in the OS (and some software). As I recall it, these were Win XP machines.
If one is using an iPad with a keyboard of some description, this is a vastly more elegant solution than dragging around a bunch of loose parts.
As you mention your family physician, why not put him on to some iPad solutions? There is a bonus of as much as $40,000 to be had for hardware/software implementations of the electronic medical records bit.
I can not recall the names of the companies, but there are several iOS developers working on apps for medical professionals. Even the VA are reported to be purchasing iPads, though the details of the intended use are still sketchy. (The VA does make extensive use of thin clients in the exam rooms which makes accessing patient records easy for the physician as well as updating those records in real time and so on. It is much more technologically advanced than many people would guess.)
Do some research and put together a presentation for you physician. Who knows you might get a consulting fee out of it and some business from his colleagues.
Cheers
@Richard, Well as I said, clumsy implementation.
Thanks for the idea for the family doctor. I’ll talk to his office manager about it.
Peace,
Gene
@Gene Steinberg, Slightly off topic, but having to do with uses for the iPad that Apple never would have envisioned, the FAA have approved the use of iPads to carry all the charts and other references that commercial airline pilots must carry with them on each flight. The brief case that each of the pilots carries weighs thirty some odd pounds. An iPad weighs nothing in comparison. Someone actually calculated the fuel savings per year from the weight reduction if you can believe that.
Here are but two of the many articles.
http://www.prnewswire.com/news-releases/united-airlines-launches-paperless-flight-deck-with-ipad-128240343.html
http://www.cnn.com/2011/TECH/innovation/02/28/FAA.approves.iPads/index.html
We do live in interesting times, do we not?
Cheers