The Apple/ARM rumor du jour is that Apple will transition its entire portable Mac line to ARM-based CPUs, dropping Intel altogether. Sources speaking to Semi Accurate claim this is a “done deal,” and the move should happen by 2013, when a 64-bit ARM A15 core becomes available. While a future generation of Apple’s A5 processor could make some sense for something akin to the MacBook Air, the claim that Apple will ditch Intel wholesale for ARM just doesn’t add up.
If the ARM CPUs are lower power, they could go massively multi-core without using too much battery. Imagine an 8-core ARM laptop.
Battery life.
I guess that running a non Intel system would put some space between them an MS, make the Hackintosh harder to make etc and possibly allow them to control and lock down the system / software more.
My guess if they go Arm laptop they would soon go Arm Desktop.
But why cant I by a powerful, Cheap Arm Netbook / Laptop which I can put Linux on – please why?
Except that Windows 8 already runs on Arm too, so a Hackintosh might still be possible if they move to Arm.
The problem with an ARM-based Hackintosh is that, since Apple has an ARM foundry, they can put whatever ARM extensions they want into their chips. There’s also very little market for ARM cores outside of the mobile market, so it’d be almost impossible to find a compatible system, with all of the features that MacOSX could be programmed to look for or that the programs would be compiled to use.
Yes you can, at least in Europe.
http://www.hercules.com/uk/ecafe/bdd/p/156/ecafe-trade-ex-hd-black-…
Because ARM cpus are anything but powerful?
Hmmm
There are many reports that Apple are going to use Intel as a foundry for their future ARM designs. I suppose that is one way to keep intel happy.
As to why Apple would do this?
The majority of laptop users (both Windows & OSX) do remarkably similar things.
1) Browse the web and use web services (eg Facebook etc) inc uploading content.
2) Use Email
3) Create simple documents.
WOT no Gamers?
IMHO they are a minority
All of those (non gamers) can be accomplished using applications that are pretty well OOTB. ie nowt special.
With this in mind, moving to ARM CPU’s is really a no brainer. Their USP would be the 24hr (or thereabouts) battery life.
With Microsoft stating that Windows 8 will run on ARM as well (it remains to be seen if that is FUD or real) also puts the death of low end Intel Powered laptops very much on the menu.
The 64K$ question is:-
Will the likes of Adobe follow and make versions of their consumer products (PS Elements) available on the new platform for both OSX & W8. If enough of them ‘Just Say No’ then using W8 or OSX on the ARM platform will have a limited life.
I can’t help wondering if the big hand of MS is muddying the waters here. They did it with Netbooks.
That said, if Apple were to release OSX on ARM there would be less chance of MS restricting the platform and keeping the WinTel virtual monopoly in place. IT all depends upon how desperate MS gets (and personally how many Linux distros release ARM versions).
As for the high end? All I can think is ‘wait & see’.
Edited 2011-05-08 06:27 UTC
That’s not going to happen. If Apple do move to ARM on the low-power laptop end, losing most major apps already in the process, they’ll use this chance to move these lower end systems over to iOS and lock them down tight. They might re-skin it and make it look like OS X, but there’s no way that control-freak Jobs would ever allow the chance for more locked-down iOS devices to get away from him. The sad bit is, most of your typical computer users simply wouldn’t care, and would probably embrace it just because it’s got a shiny Apple logo on it.
I greatly doubt that control freak Jobs will still be around in two years to influence anything.
But why do they really need it ^A'locked down tight^A'?
Simply by shifting the low-end ^A'iBook^A' market to ARM itself is a disruption that will almost certainly be able to boost Apple^A's App Store. Many developers will CHOOSE to only distribute their ARM versions via App Store… Once there is inertia, ANY ARM App will by default be offered on the App Store (if not exclusively) once that is what consumers expect, and Apps that DON^A'T do this will tend to lose market-share to the convenience of App Store downloads. No need to ^A'force^A' anything.
What does getting 100% accomplish for them? Forcing developers who don^A't want to / can^A't conform to App Store rules to no longer offer their product to Apple^A's consumer majority? I don^A't think Apple REALLY want ANOTHER OSX platform to worry about, keeping it simple with iOS, iOStablet, and OSX (x86 and ARM) is very manageable, the different executable format is trivial.
Such a shift also gives Apple better differentiation vis-a-vis their Pro-line, which right now isn^A't really that differentiated: basically screen size (which is totally artificial), and I/O options, which most of the Pro-market (people buying them) doesn^A't care all that much about (people who want 15^A¨ screens are probably equal to those who need hi-speed/expandable I/O). With a more differentiated ARM low-end, that let^A's Apple much more organically differentiate the Pro-line, e.g. for people who want real x86 performance… An obviously differentiated successor to the 12^A¨ Powerbook is now realistic, and Apple can now sell 15^A¨+ iBooks to more low-end customers for more net profit than the people forced to get MacBookPros for a large screen.
Edited 2011-05-08 15:06 UTC
Disagree. Macs are also frequently used for multimedia work (it is said that OSX does a pretty good job at handling low-latency audio and such, but it’s probably more of a cultural thing). Is ARM ready for handling heavy loads and unoptimized monsters like Cubase ? This remains to be seen…
Edited 2011-05-08 12:24 UTC
Yes, that is true. Standard Windows drivers are awful in comparison to Core Audio. So – same hardware (MacBook 2007 model, 4GB RAM, Mac OS X 10.6.6 vs Windows 7 professional 32bit) using Reaper under OS X with no additional drivers, perfect flawless low latency audio recording for 5 tracks built up in turn over a session (recorded one at a time) However, Windows, we have synchronization after the first track with standard Windows drivers. Installing ASIO4ALL (ASIO driver emulation) we get passable results. Conclusion – Mac OS X is better at low latency audion “out of the box.”
Just to clarify – same with 10.5 (where it’s running in more 32bit mode – though the version of Reaper is 32bit anyway) and this is using the Mic jack built in to the hardware with no external USB/Firewire sound cards included.
Edited 2011-05-09 11:19 UTC
Apple have switched to ARM.
iOS devices are their biggest profit area. Macs only represent a quarter of the business.
I’m all for that. Intel’s boring. Although I’m not so hot for Apple right now (TROLL!!! + suing binge), should they go this route, I’m cheering 100%. Shall I get the pom-poms?
Intel performance per clock is impressive. However we currently have a 95W cpu standard to show for it (and don’t even get me started on gpus). A move by apple theoretically could drive a new standard…ie to 9.5W (a factor of 10 less). I definitely would go for that. And for 95% of users this would definitely do the trick.
Currently the A9 @ 500Mhz is about the same performance wise as an Atom @ 1.3Ghz.
If you look at the fact that the A9 uses 25% of the wattage of the Atom; you can easily double your batter life. Remember that the main power consumers are the CPU, Disk, Screen.
MS Has already ported Windows 8 to ARM. They’ve also asked the ARM people to develop a 48 core ARM system for the server room.
Apple is already working on reducing the number of supported CPUs for Mac OS X. 10.6 dropped PPC support and 10.7 is dropping 32bit support (not this just means the ability to run the OS on a 32bit Intel system not the support to run 32bit apps). Also 10.7 is dropping PPC emulation.
iOS is just Mac OS X in a limited state. Apple is currently working on merging the iOS & Mac OS lines.
Apple is heavily invested in improving the compilers for ARM. Apple is also on of the founding members of the ARM group.
I don’t see any reason you couldn’t run Windows 8 ARM on an ARM based Mac.
The biggest loser on the move to ARM will be Intel & MS. Intel may build chips ARM chips under contract; however, the chip design will be done by others. The x86 market will also start collapsing which means that they’ll need to move into high-end servers and hope they can stay ahead of ARM development.
MS will have lots of issues. One, is getting people to port their applications to ARM; this will be somewhat mitigated by the fact that they are currently telling everyone to do .NET instead of Win32 development (.NET is a virtual processor so any .NET application will just use the JIT and run normally). Two, MS will finally need to break backwards compatibility; 16bit DOS & Windows programs will be lost. A third issue is the lack of support for multiple targets in a single binary; thus, vendors are going to have to ship 2 versions of all applications; one for x86 and one for ARM.
My research area focuses on CPU architectures and energy reduction, among other things. For certain areas, we tend to take an optimistic attitude that most algorithms will be parallelizable, so it’s better to slow down cores for energy efficiency and then go wider for throughput. Seriously, if you were to run a 1V CPU at 400mV, you’d get something like a 100X power reduction, but you’d also make everything run 10X slower. For any work, you get a 10X energy reduction, and all at much lower power.
Unfortunately, not all problems are so easy to parallelize, and so practical systems still rely heavily on single-thread performance, with architectures still focusing on instruction-level parallelism (out-of-order and superscalar).
I’m excited about the A15, and I’m sure that if someone were to use it in datacenters, then although latency on a lightly-loaded system would go up (a tad, since the network is usually the major source of latency), the system would be able to handle a hell of a lot more concurrent users for the same power budget. But the fact is, ARM chips are designed for energy efficiency, not performance, and for a very long time, Intel’s best is going to vastly out-perform ARM’s best, when you ignore power per thread.
People buy Macs for lots of different reasons, but those buying high end machines (like my 17″ MBP) are expecting performance. Apple knows this, and this is going to keep them from switching to ARM in their mainstream personal computers.
That being said, in theory, if the MacOS core were build for ARM _and_ x86, Apple could sell machines that married an ARM with an x86. The ARM would be used for OS and non-performance tasks, while the x86 would be used for apps and perforance stuff. Using the ARM more would allow the x86 to sleep more, saving gobs of energy. But I doubt Apple would do it because of the complexity it would add.
One final note: Be careful when you talk about power efficiency and energy-efficiency. The most power-efficient device is a device that’s turned off. It just takes forever to do anything, so its energy efficiency is horrible. What you really want is more battery life. Batteries store (potential) energy, in the form of electric charge. When researchers and engineers optimize for efficiency, they may optimize for power, which is to reduce power, holding runtime constant. Or they may optimize for energy. Or they may optimize for the Pareto surface of runtime vs. power or energy. Optimizing for Energy often means increasing runtime (for a much greater reduction in power). But when runtime is important, sometimes we optimize for the fictitious units Energy*Delay or Energy*Delay^2. BTW, Power*Delay^2 (or Energy*Delay) is a relatively voltage-invariant metric, since power (at constant frequency) goes up with the square of supply voltage, while frequency goes up linearly (give or take), and frequency is the reciprocal of delay (not counting the memory hierarchy delays).
You think of x86 as performance. But it is really an awkward arch that was an underdog until recent years.
And it was easily beaten on performance side by cheap and simple console cpu arch (CELL) even with all microarchitecture and manufacturing process advances what cost Intel/AMD an astronomical amount of money.
Apple have never been interested in the lower end market of things.
You can pick a laptop up for say $400 and it’s a cheap pile of rubbish, but it will browse the Internet and do emails no worries.
Apple laptop on the other hand is $1400, but at least it’s going to last you a good 4 years I’ve found and it’s not going to struggle if you have word an do some light photoshop work and process your photos and make a movie.
If the only benefit is battery life, then a lot of people will leave. Who cares about battery life if you can’t do your work properly.
Now if for some reason they can pull of battery life, and better performance than Intel, sounds good. But ARM better have a damn good road-map for constant improvements and the ability to manufacturer enough of the chips otherwise there’s going to be a lot of pain.
That was one of the reasons Intel won out over AMD. They had a solid road-map with exactly what they were doing and where they were going. And their market share and performance have shown that.
ARM’s manufacturing ability (or lack of) is irrelevant.
One of the reasons Apple stated for changing from PPC to x86 was power use, so if they still find this important, ARM is a logical choice for a new transition, now that they are powerful enough to power laptops and servers. Also, Apple has greater experience than most in changing processors: They started the Mac series with 68K, then switched to PPC and later to x86. They have now used x86 longer than either 68K or PPC, so you could argue that they are overdue by now.
But the main point about this is that Apple have become increasingly careful not to rely too much on a specific processor architecture, so the cost of changing is far lower for Apple than it is for Microsoft, who nevertheless will make Win8 run on ARM.
These days, most of the heavy loads on a computer are run on graphics processors and other coprocessors, so the performance of the main CPU is less than it used to be. It is far more important that you can get the coprocessors you want. Here ARM has an obvious advantage over x86: It is fairly easy to get an ARM license to produce an ARM core surrounded by whatever coprocessors you like, an option Apple took for its A4 chip. With x86, you have to take whatever SoCs Intel, AMD and a few other make.
As for running “old” code, Apple showed with its previous transitions that emulation isn’t a problem: Most time is spent in library routines, which can run natively. And with JIT, emulation of the rest can be done with relatively low overhead (around x2). ARM and x86 share the use of arithmetic flags, so emulation of flags, which is a problem when emulating x86 on, say, PPC or MIPS is almost a non-issue when emulating on ARM. In the future Apple will ship more and more code in the LLVM format, which is compiled at load time to native code, so they can ship the same code to both x86 and ARM-based computers.
As other’s have mentioned, using a SoC of their own design and under their complete control, Apple can prevent porting of future versions of MacOS to non-Apple machines. As Apple makes its money from hardware sales (at least with the Mac line), this is something they are eager to do.
The main problem with ARM so far has been performance and support of more than 4MB of memory. The A15 core solves both of these problems, even if it isn’t a true 64-bit core. But 64 bits is mainly an issue for memory — the performance gain is modest. For x86, the main performance gain from 64-bit processors is due to a different FP instruction set; the integer performance isn’t that much greater.
Tiny nit – They used PPC from 1994 until 2005, that’s 11 years. They’ve had x86 from 2006 until today, only 5 years. If you mean exclusively PPC, That was starting with MacOS8.5, released in 2007, so even then it’s 8 years. Oh, and 68k they used for the 10 years prior to PPC.
People actually do believe that other people are stupid and will leave the power, cheap cost of x86 + Windows to go on ARM with OSX/ios? To do what? Make friends on facebook and sending twits? They can do that wit a phone, they don’t need a computer for that.
Apple is using ARM only for low power devices when they need to optimize for battery life.
And guess what? You can’t use Photoshop or play Crisis on iPhone, Ipad, Ipod or any other iCrap devices.
Even if MS will make Windows 8 ARM compatible, I doubt that you can use that for Photoshop, video transcoding or doing any serious stuff which requires computing power.
I understand that lots of people hate Microsoft, Intel and whatever, but guess what x86 is still the most powerful CPU architecture and Windows is still the most used OS, if not the best. And that will be the case even after 10 years from now.
Saying that Cell, Power PC, MIPS, ARM are better than x86 is also FUD and saying that only demonstrates that you are a computer illiterate. Supercomputers of today are made from, Nvidia and AMD GPUs, and guess what, x86 CPUs.
Please take a look at http://www.top500.org/
Note that 2 of the top 10 are running Power/PPC.
x86 is just cheap and easy to get. The Intel/AMD based systems are running close to 200,000 cores.
x86 didn’t get any major performance improvements until they replaced the back-end with a RISC core and did a lot of work on the front-end to recompile the x86 instruction set into a RISC instruction set.
Look at the Intel/Atom. This is the low power version and to meet power consumption requirements, they had to simplify the x86->RISC process and remove out of order execution. The performance on this chip is bad and it still runs hot.
Microsoft is using PPC for the xBox systems for performance reasons.
Apple dumped PPC only because of a lack of a road map for Desktop processors. PPC today is used mostly in servers and embedded systems.
Your perceived lack of performance on the ARM chips is mostly caused by not know what the ARM market is. ARM was created for and is mostly used for embedded use. Thus, power usage is more important then speed. The A9 ,at half the speed of the Atom, is still faster; there is no reason why you couldn’t clock an ARM at the same speed as an x86 process. The overclocked ARM would still use less power and give you much better performance when compared to x86 chips. If you want even better performance in ARM, you could always add in some of the x86 type tricks to speed things up; of course doing this will increase the power consumption which is why ARM hasn’t gone this route.
Nope, is easy to get better CPUs. It’s just that WinTel is the major player in the desktop market. Just for reference the ARM market is much bigger then WinTel.
Photoshop7 worked smoothly on Pentium MMX class hardware.
ARM can run circles around that setup.
Video transcoding, etc can be done with NEON, or better with GPU.
BTW, PowerXCell-8i achieved 200GFlops/s (single-precision) and 100GFlops/s in double precision.
Which intel cpu is capable of this? And when it was released? First CELL was manufactured in 2005.
Edited 2011-05-09 17:37 UTC
All this may simply depend on if Steve stubs his toe tonight.
It’s happened before.
“rumor du jour”? Come on, the first time I read this was even before the x86 switch
Seriously, this MAY happen in the future, but not in 2013. A lot of people (and I mean A LOT) is happy to buy a Mac nowadays because it can turn into a Windows machine in a few clicks.
There are consultant who keeps a distinct Windows virtual machine for each customer, while doing their stuff on the underlying OS X. We’re talking enterprise here – and you know how slow enterprises are to move to new tech… I still see Win2k on corporate desktops and servers everywhere.