Apple Considering Switch From Intel Chips in Future Macs

[threesixty360]

missing the point...

I think alot of us are missing the point again.
ARM's main advantage is CPU power or battery life (both of which I'm sure
Intel can beat in the long run). ARM's main advantage is self determination.
Most of Apple's issue's in the past have come from their reliance on partnerships.
At one stage or another they have relied on Microsoft for Office, IBM for chips, Google for maps,
Adobe for flash/photoshop etc...

And they DONT LIKE IT.

It makes them vunerable and it means they dont control their own destiny.
The real use of all these profits and the long term goal of Jobs is for Apple to control
everything they need or design things in such a way that they have options. Options allow
you to bargain with vendors and get the best deal. You cant be beholden to a supplier.

Look at the PC market, Dell, HP, Acer just had to sit their and watch MS announce that they were
entering their markets. All the OEM's are stuck with Intel and MS and they cant get out. They
cannot determine their futures. If Intel or MS mess up or are not aligned with their vision the OEM's
are finished.

ARM as a framework allows Apple to control their destiny. It also allows them to differentiate
their products and make them unique in the market. If they decide to put emphasis on GPU power
in their SOC, they can. If dont want to include NFC they can.THey can do what they want. You
cant always do that with Intel.

The reality is that most of the money in selling devices is coming from the low powered end of things.
Apple are pushing that side and are doing well. Look at how people were surprised with A6.
I can see Apple moving into the cheaper laptop segment with 499 laptops powered by ARM. But
it will be the best low powered machine around and optimized as such. And it will run apps
from the App store that all they need is a recompile. I doubt it will be iOS based because that
will require touch screen input and Apple arent going to merge the two.

Steve Jobs once said that they dont know how to make a cheap PC laptop well. SO they made the ipad.
Now after making that, they KNOW how to make a cheap pc. ARM is the key. And they'll finally
do it.

 

[ChristianVirtual]

In addition, to compete with Intel and Samsung, Apple may have to produce his own CPUs, so is a purchase of TSMC possible?

I'm sure Apple could buy a chip plant from Motorola for few bucks ... Google wanted only the patents ...

 

[toke lahti]

Which one have more horsepower: G4 or A6X?

 

[KnightWRX]

I think alot of us are missing the point again.

Quite the contrary. Most of us get the point. The problem is that we don't like where this is going. Sure Apple is going to be massively more profitable, but in the end, to us, the people discussing in this thread it makes the Mac less useful. That it may sell more to other types of users out there is besides the fact that to us, it may mean going to other vendors.

ARM has no advantage over x86, in a technical sense. Like you state, the advantage is purely for Apple's bottom line. But their bottom line means moving away from a certain group of users to another group of users, and frankly, no one likes being dumped after being milked for years.

Anyway, this should be no surprise. A guy once said the following, in the 90s :

If I were running Apple, I would milk the Macintosh for all it's worth — and get busy on the next great thing. The PC wars are over. Done. Microsoft won a long time ago.

As quoted in Fortune (19 February 1996)

Prophecy ?

----------

I'm sure Apple could buy a chip plant from Motorola for few bucks ... Google wanted only the patents ...

Google did not buy chip plants. They only bought Motorola Mobility.

 

[ChristianVirtual]

Which one have more horsepower: G4 or A6X?

G4: http://browser.primatelabs.com/geekbench2/view/480496

So A6X wins ...

 

[KnightWRX]

Which one have more horsepower: G4 or A6X?

The Qualcomm Kraits which are comparable to the A6X are rated for about 10,000 MIPS (Million of Instructions Per Second). Compared to the POWER architecture chips, that would be equivalent to 1 of the execution units found in the PS3 (the PS3 has 6 + 1 execution units if I'm not mistaken).

That puts the A6X capabilities at around 2003 levels for desktop computing.

 

[Neodym]

That's pros for Apple, not for consumers. I am a consumer, why would I care about that? What would *I* gain from Apple going Arm?

I interpreted your original question as to why Apple would make such a move. You could profit (depending on your individual needs):

• Because Apple stays alive in a changing market
• From lower prices or a better price/performance ratio
• From being on the forefront of technology if the market swings to another major hardware platform
• From solutions possible due to new technology (massively multi-core) that are not possible with x86 right now
• From fast product updates due to lower dependency on 3rd party suppliers for crucial system components
• From lower energy consumption (= longer battery time)
• If nothing else from the competition Intel gets from ARM, leading to lower prices and/or better price/performance ratios on x86

Because most software is not written in a way to run in such a parallele execution environnement, just putting in more cores does not give you the equivalent of an Intel solution.

You assume that this would not change. A few years ago people questioned even dual-core architectures - nowadays quad-core is considered standard and more and more software comes to support this.

And a 32 core ARM processor, on top of being less able to perform, wouldn't be more power efficient than an x86 quad-core or even dual-core machine.

The current Mac mini contains CPU's with up to 45W TDP. A Cortex A9 DualCore has a TDP of only 2W. That would allow for roughly 20 ARM Dualcore CPUs = 40 cores. Granted - you would need dedicated software to make use of that massive core count.

But if Apple would provide necessary tools, then maybe it would take much less than 40 cores working actively for similar performance - and then you would see power savings indeed. Or you may see higher performance for the same power consumption. Current iPhones offer a performance that was available in iMacs only a few years ago - and development on the ARM platform is blazingly fast...

There is nothing to gain here except being different. Which is dumb. Change just to be different brings nothing good, especially when it hurts with compatibility.

I do know managers who consider more proprietary designs as better as they can sell more of their own products. This may even work if you are market leader.

The situation here is a little different though: As Windows is also headed for ARM, that may be the next major hardware platform and Apple as well as Microsoft (and others) would be highly interested to be there in time to set the standards and earn the money from being first (or at least first follower) on the market.

So there may be quite something to lose here, but at least as much to be gained - perhaps significantly more. In other words: "The risk of going for that market is lower than the risk of NOT going for that market."

Many gaming titles on Macs these days exist just because the Mac is x86 and software porting houses use the Windows code base wrapped around with a specific WINE build (Guild Wars 2, Batman: Arkham stuff, etc..). Forget that with ARM Macs.

Apple never considered their desktop computers as gaming platforms - else they would have put more effort into the GPU, which they didn't. On the other hand a move to ARM would open the complete iOS gaming market. Sure - right now you don't see complex shooters, RPG's and similar stuff on iOS, but that platform evolves quickly and the desktop gaming market won't be static as well.

Same for all the virtualization stuff (VMWare, Parallele, Virtual Box).

Sorry, but that's plain wrong! There has been virtualization stuff in the past across all hardware boundaries. Surely it's easier to simply write a small wrapper which hands over the commands directly to the CPU. Surely an emulation will require significantly more power in comparison. But it _can_ be done. And with a higher core count it may reach similar performance levels to today's approaches - unlike to former emulations.

In former times you had 1-2 cores having to emulate 1-2 cores of a completely different hardware (like e.g. PPC and x86). Which resulted in a significantly lower performance of the emulation compared to the original.

Now imagine you would have 16 (ARM) cores to emulate a 4-core CPU of a completely different platform (like x86) and still have 8-16 cores left for tasks of the host OS...

Forget Bootcamp and running Windows 8, you'll be stuck with Windows RT which is not compatible with the full suite of Windows software out there.

As i said above - there are other ways, which may require some more effort, but they are perfectly possible.

It would take a joint move from both MS and Apple to make it work.

Well - if both Windows and OSX switch to ARM, that's about 98% of the desktop OS market. I doubt many developers would refuse to follow that route - else they would simply vanish from the market...

 

[ChristianVirtual]

Google did not buy chip plants. They only bought Motorola Mobility.

Opps, my bad ... thanks for clarifying ...

 

[adder7712]

Now this one's going to need a bit of explaining.

Indeed. I still can't fathom how a dinky quad core Cortex-A9 compares to a "crap" Intel Core i5.

 

[honam1021]

Way too early, if they switch to ARM it'll make macs even worse in gaming

 

[threesixty360]

Quite the contrary. Most of us get the point. The problem is that we don't like where this is going. Sure Apple is going to be massively more profitable, but in the end, to us, the people discussing in this thread it makes the Mac less useful.

Lets be honest. The mac with x86 is overpowered for a vast majority of users need. Apple manage to sell machines at these prices because of huge brand loyalty and commitment to design. However that position is an exceptional one as no other OEM can sell pc's in volume at the price point. The writing is on the wall, they cant keep selling computers at the those price points.

With win8 and even Android with the chromebook we will so computers that look like macbook air's but will be half the price. If apple dont prepare for that eventuality they will lose the PC market all together and just have iOS. I think on the contrary, moving some of the Mac line to ARM is in a Mac users interest because it will preserve the platform and keep it growing.

Apple themselves use these machines to carry out their business tasks. So what you see with Apple is a reflection of whats not just profitable, but what is important to them as workers. The mac pro is not important for most engineers and designers / developers in 2012, hence less love. Light weight, low powered computing is more important to them and frankly, everyone else in the world.

Their will always be high powered machines from Apple because they need them to do their jobs. So I dont think you have to worry about that.

 

[XboxMySocks]

Haha! Apple might as well put a gun to its head. My Pentium 4 from eight or nine years still has a good edge over any mobile CPU. According to Geekbench, the A6X manages to get an impressive score of around 1700. Compare that to the Mac Pro though, which benches scores of over 38000, and I think you can all see a problem. This forum in particular likes to say specs don't matter, but I don't think even you all would be stupid enough to fall for a move like that.

Wait what? Are you seriously suggesting that a phone should benchmark as much as a full blown computer should? lol wtf

 

[KnightWRX]

You assume that this would not change. A few years ago people questioned even dual-core architectures - nowadays quad-core is considered standard and more and more software comes to support this.

I assume nothing. Parallele programming has been with us for decades. It was a hard problem 30 years ago, it's still a hard problem today. GCD makes it a bit easier to run tasks asynchronously, but in the end, you have to divide your tasks, and you might just not have enough tasks to spread out in the first place to take advantage of all those cores.

Multiple cores does not replace a single fast core.

The current Mac mini contains CPU's with up to 45W TDP. A Cortex A9 DualCore has a TDP of only 2W. That would allow for roughly 20 ARM Dualcore CPUs = 40 cores. Granted - you would need dedicated software to make use of that massive core count.

My point. Proven.

But if Apple would provide necessary tools, then maybe it would take much less than 40 cores working actively for similar performance - and then you would see power savings indeed. Or you may see higher performance for the the same power consumption. Current iPhones offer a performance that was available in iMacs only a few years ago - and development on the ARM platform is blazingly fast...

ARM has no energy efficiency built in. ARM is not moving any faster than x86. Don't mistake instruction sets with chip design.

I do know managers who consider more proprietary designs as better as they can sell more of their own products. This may even work if you are market leader.

Sure, that's the whole reason Apple would go with their own ARM design rather than x86, and they wouldn't give a damn about comptability. But you know what ? I don't give a flying ... about their bottom line, if their stuff isn't able to run what I need, I'll have to go to another vendor.

I don't have to like it.

The situation here is a little different though: As Windows is also headed for ARM, that may be the next major hardware platform and Apple as well as Microsoft (and others) would be highly interested to be there in time to set the standards and earn the money from being first (or at least first follower) on the market.

Except there is no indication that Windows is going to ARM. Microsoft is pulling quite the marketing trickery and some of you guys are believing it. Windows RT is not Windows.

So there may be quite something to lose here, but at least as much to be gained - perhaps significantly more. In other words: "The risk of going for that market is lower than the risk of NOT going for that market."

The gain is not for the user, it's for Apple. Don't mistake it.

Sorry, but that's plain wrong! There has been virtualization stuff in the past across all hardware boundaries.

That's called emulation, not virtualization. The small "wrapper" you refer to is what is known as an hypervisor. Don't try to educate me on this topic, you have probably much more to learn here.

The problem with emulation is that it comes at the cost of performance. When your top of the line chip can only perform at 2003 levels, you're not going to emulate anything at 2012 levels anytime soon.

Now imagine you would have 16 (ARM) cores to emulate a 4-core CPU of a completely different platform (like x86) and still have 8-16 cores left for tasks of the host OS...

The problem is that again, parallele programming is not as easy a task as you make it sound, and you can't replace a single fast core with multiple slower cores. Especially in an emulation scenario. Emulation does not remove the need for instructions to run off the core or for mutexes or other dependencies between your instructions.

Well - if both Windows and OSX switch to ARM, that's about 98% of the desktop OS market. I doubt many developers would refuse to follow that route - else they would simply vanish from the market...

Developers haven't targetted CPU instructions sets for years. That's not the issue at all. Developers mostly write their code to higher level APIs and frameworks if not for VM or interpreters that are mostly portable and multi-platform already.

If you really think that is the issue, you're sadly mistaken.

----------

Wait what? Are you seriously suggesting that a phone should benchmark as much as a full blown computer should? lol wtf

If someone attempts to put that phone CPU into a laptop or desktop and sell it to you as a bonafide Mac, yes, it should.

 

[pubwvj]

I wish that Apple would support emulation for all past Macintosh software all the way back to MacOS1.0. Heck, they should go all the way back to the AppleI. There is a tremendous amount of educational software that was created during the 1990's that has never been redone for Intel and MacOSX. It used to run under Classic but Apple abandoned it. They are destroying both cultural heritage and educational resources. They have the money to keep up the emulation and it would vastly expand the media available to run on their machines which would make more people interested in upgrading to the latest and greatest hardware thus promoting more Apple sales and more money for Apple's pocket. Heck, they could even offer full Windows, DOS and CPM emulation and take over the whole market.

 

[KnightWRX]

Lets be honest. The mac with x86 is overpowered for a vast majority of users need.

I said the very same thing in this very same thread. But again, you're missing the point : The users Apple would leave behind with such a move are the ones that are discussing things here and now in this thread, not the masses that don't need the power of x86 chips.

----------

I wish that Apple would support emulation for all past Macintosh software all the way back to MacOS1.0.

That's just asking for problems, cruft, performance issues, security holes, etc.. You've already been told your wish is about the worse thing that could happen to computing. Ever.

Get over it, luckily for the rest of us, you're never getting it.

 

[MH01]

Steve Jobs once said that they dont know how to make a cheap PC laptop well. SO they made the ipad.
Now after making that, they KNOW how to make a cheap pc. ARM is the key. And they'll finally
do it.

They will not make cheap PCs, the margin will just increase.

 

[VulchR]

Lets be honest. The mac with x86 is overpowered for a vast majority of users need....

It's not for mine, or most researchers using macs in STEM in academia. Right now Mac's are a fantastic choice for technical work because they run OS X, 'UNIX' (close enough for my use) and Windows. Any move to a sand-boxed (or heaven forbid, cloud-based) laptops or desktop machines based on a CPU that prioritizes power consumption over performance and compatibility will kill of much of Apple's presence on the world's university campuses, and thereby reduce the numbers of the next generation of Apple's customers. I am not involved in the creative industries, but I would imagine the same holds true there as well.

Apple's current success is about the ecosystem. An essential part of that ecosystem is reasonably high-performance desktops/laptops that are more flexible than any other computers available. Anything that even hints at endangering that will see Apple's market share dry up inexorably. I hope Apple is not so stupid and foolish as to go down this road. It leads nowhere.

EDIT: I get that the article talks about laptops, but I can't imagine Apple's whole line up would consist of ARM machines with the exception of high performance Intel laptop/desktops. I'd be willing to bet they migrate their whole line to ARM if they went this route.

 

[slughead]

There's just no way.

 

[KnightWRX]

Apple's current success is about the ecosystem. An essential part of that ecosystem is reasonably high-performance desktops. Anything that even hints at endangering that will see Apple's market share dry up inexorably. I hope Apple is not so stupid and foolish as to go down this road. It leads nowhere.

You're mistaken about one thing. Apple's market share won't dry up with iOS like devices completely replacing Macs. It'll just shift to a more consumer market and away from more computing intensive environnements.

Big gaming titles, virtualization, scientific computing, media content edition and creation, it'll just go away for Apple and Apple will simply sell "lifestyle" type devices, aimed at consumers wanting to do the same simple tasks they were doing 10 years ago, on now more efficient devices that are smaller and more out of the way than ever.

That roads just leads Apple deeper into the mass market consumer segment and farther away from its current base of "geeky" computer users that like OS X because it's Unix, but at the same time, it's convenient for everyday use.

 

[Sensation]

Silly idea, why put crappy Arms in desktops. Make all apps incompatible all over again and confuse the users.

 

[paulsdenton]

The governor used to say...

One of the best governors Vermont ever had, Howard Dean, used to repeat an old Green Mountain State aphorism:

"If it ain't broke, don't fix it"

I think that might apply in this case!

 

[AppleMacFinder]

I kind of hope you're trolling here, as this doesn't even resemble accurate, and it doesn't sound like sarcasm, given that you cited a brand in there. Ghz are a poor measure of overall performance without further context, regardless of architecture. The only area where they're still used is in differentiating between two very similar cpus. Intel's 17W cpus would still blow away the 2W ones you mentioned. While the ARM versions could sacrifice performance in favor of that 16W, they don't represent the entire power budget for the machine. If you aren't trolling, you should catch up on your reading. You're making a nonsense comparison when Intel designs cpus anywhere from sub 10W with Atom to 130+ for servers and workstations.

My point is that ARM CPUs have a much higher overall Performance/Power ratio (not only in GHz terms),
that is why they are the future. And Intel Atom is a joke, because its perfomance (and also the performance of integrated graphics) is terrible.

 

[Thunderhawks]

I wish that Apple would support emulation for all past Macintosh software all the way back to MacOS1.0. Heck, they should go all the way back to the AppleI. There is a tremendous amount of educational software that was created during the 1990's that has never been redone for Intel and MacOSX. It used to run under Classic but Apple abandoned it. They are destroying both cultural heritage and educational resources. They have the money to keep up the emulation and it would vastly expand the media available to run on their machines which would make more people interested in upgrading to the latest and greatest hardware thus promoting more Apple sales and more money for Apple's pocket. Heck, they could even offer full Windows, DOS and CPM emulation and take over the whole market.

Living in the past may be romantic, but unrealistic.

Everything is driven by money and there is no money in education, otherwise it would be better.

It is not Apple s job to worry about cultural heritage. Their job is making money.

As for educational resources I can guarantee you that ANYBODY who needs to research something will find a way to do it with a little effort.

Besides that what is wrong with books? They are still around last I checked.

There is all your heritage.

Unfortunately today s generation has no attention span to read. If it can t be clicked on and something plays they give up.

Come into this century as messed up as it is

 

[KnightWRX]

My point is that ARM CPUs have a much higher overall Performance/Power ratio

I have seen nothing of the sorts. ARM CPUs have a power advantage because they sacrifice sheer number crunching ability.

 

[iBug2]

That roads just leads Apple deeper into the mass market consumer segment and farther away from its current base of "geeky" computer users that like OS X because it's Unix, but at the same time, it's convenient for everyday use.

Geeky computer users are a very small minority of Apple's OS X user base. I don't think Apple would really care about those. But they should care about the userbase of desktop publishers, audio/music editing, video editing, photography etc. Even those people don't make up the majority of their OS X userbase but they are a much bigger group than us geeks.