backtop


Print 50 comment(s) - last by ClownPuncher.. on Nov 7 at 2:47 PM

Apple may co-design special new ARM chips for its laptop and desktop computers

Bob Mansfield, Apple, Inc.'s (AAPL"Technologies" group senior vice president, recently announced that his semiconductor teams have "ambitious plans for the future."

I. Change Ahead?

For Apple, who has been quietly been growing its chipmaking army with acquisitions like P.A. Semi and Intrinsity, that could mean a switch from Intel Corp.'s (INTC) x86 chips on the Mac computer front to in-house system-on-a-chip (SoC) designs utilizing ARM Holdings plc.'s (LON:ARM) proprietary architecture, which is currently found in Apple's iPhones, iPods, and iPads.

new report from Bloomberg suggests Apple is exploring a possible exit from its dependence on Intel chips, over the next several years.  The news agency writes:

While Apple is now committed to Intel in computers and is unlikely to switch in the next few years, some engineers say a shift to its own designs is inevitable as the features of mobile devices and PCs become more similar, two people said.
....
As handheld devices increasingly function like PCs, the engineers working on this project within Apple envision machines that use a common chip design. If Apple Chief Executive Officer Tim Cook wants to offer the consumer of 2017 and beyond a seamless experience on laptops, phones, tablets and televisions, it will be easier to build if all the devices have a consistent underlying chip architecture, according to one of the people.

 
ARM core
Apple is rumored to be contemplating switching its Macs to ARM chips.
[Image Source: Maximum PC]

None of the companies involved were willing to comment on the rumor.  However, this has not been the first time that Apple has been rumored to be working on computer-side CPUs in-house.

ARM shares climbed 4.2 percent on the London Stock Exchange following this news, and news that ARM had paid MIPS Technologies, Inc. (MIPS) a lump sum of $350M USD to license its portfolio of 580 hardware patents.  The MIPS deal lays to rest concerns that the rival reduced-instruction set computer (RISC) architecture designer might be preparing to sue its more prolific peer (ARM).

II. Lesson From the PowerPC Era

Apple may be reticent to jump into a computer-side deployment of ARM too soon.  

Only a decade ago Apple was supporting another non-x86 architecture -- PowerPC -- whose chips were co-designed by Motorola and International Business Machines, Inc. (IBM).  While PowerPC Macs seemed to be faster in some benchmarks than their higher-clocked Windows/x86 counterparts, Intel's flashy clock speeds during the height of the Megahertz wars hurt Apple from an image standpoint.  And Motorola's decision to bail on PowerPC drove Apple to in 2005 finalize its long-standing plans to jump ship to Intel's x86 chips.

ARM is hardly analogous to PowerPC at present, but it's easy to see why Apple might be uneager to move too fast, should market conditions change.  Indeed, the next few years in the chip market are a time of great uncertainty.  

PowerPC G3
Apple's decision to ditch PowerPC brought it closer in line with Windows OEMs, who also use Intel chips. [Image Source: Bryon Realey/Flickr]

On the one hand, Intel is pushing hard to leverage its process advantages to deliver fast, power-efficient mobile cores, cutting into ARM's core business.  On the other hand, ARM is invading the personal computer market with Windows RT, and is expected to land in the server market around 2014 with Window Server 2012 R2.  ARM has been boosted by x86 veteran Advanced Micro Devices, Inc.'s (AMD) decision to at least partially transition its CPU line to ARM chips.

If Bloomberg's sources are accurate, Apple will likely be hard at work designing a possible ARM-based SoC replacement to drop into its MacBook Airs and MacBook Pros.  But the decision to pull the trigger may ultimately rest on where ARM stands versus Intel/x86 come 2016-2017.

Source: Bloomberg



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Very ambitious.
By JackBurton on 11/6/2012 1:31:02 PM , Rating: 2
That's a VERY ambitious undertaking. However, if there is anyone that can pull it off, it's Apple. Another advantage of controlling the software AND hardware.




RE: Very ambitious.
By XZerg on 11/6/2012 1:42:30 PM , Rating: 1
ditto and i can very much see Apple going for this. this would add one more level of conformity to a "standard" across the Apple products going forward.

With the world becoming based on more and more cross platform compatible the cpu architecture becomes less and less relevant. Before you underestimate this statement imagine the following:
A server computer at home/work hosting compute power and rest of the systems are then just dumb terminals that consume/"produce" content leaving the grunt work to the server to perform remotely. If that's the case, does it really matter what architecture the dumb terminals use? No.

Google and Apple are two companies that can proceed with this concept. Microsoft is getting there and Windows 8 is just what the doctor ordered - common GUI across two major consumer platforms.


RE: Very ambitious.
By amanojaku on 11/6/2012 1:52:19 PM , Rating: 3
What neither of you understand is the design goal ARM has: low power consumption first, performance second. Apple won't be able to refine an ARM CPU to the point that it can compete with x86 performance unless it abandons low power consumption as its primary target. If the source claimed this was just for Mac Airs I could see this working. For a full-blown laptop not even a 64-bit ARM can compete.


RE: Very ambitious.
By XZerg on 11/6/2012 2:13:32 PM , Rating: 2
did you read the part about dumb terminals? do you honestly feel that most of the smartphones and tablets out there today cannot take the place of the dumb terminals? they can easily do so. Hence, switching the architecture would do absolutely zilch in that ecosystem.

I, myself, have been looking to do that exact setup for years at home. Yes there are some applications the technologies/ecosystem are not mature enough for at the moment but if a company as big as Apple/Microsoft/Google were to push it then it could become the next generation computing. These three have the might and know-how to do so too, unlike many of the computing devices makers only.


RE: Very ambitious.
By amanojaku on 11/6/2012 3:05:06 PM , Rating: 2
Since when does Apple sell dumb terminals? And Apple abandoned the server market a few years ago. Your idea makes sense for any company other than Apple. The source mentioned Airs and MacBooks, neither of which would be satisfied with a dumb terminal processor. Other companies are already making dumb terminals with ARM CPUs. Look at Wyse.


RE: Very ambitious.
By someguy123 on 11/6/2012 3:36:51 PM , Rating: 3
That's not really a massive dependency overhaul like what the article implies. Basically they could just turn their apple TV into a dumb terminal, which already uses an A5.

I really don't see apple making a complete transition to ARM unless ARM can prepare some high performance designs. PPC risc processors made sense because PPC was a decent amount faster than intel's x86. ARM on the other hand is all performance/watt and would never keep up at all, especially with apple going resolution happy with the macbooks.


RE: Very ambitious.
By 91TTZ on 11/6/2012 4:30:29 PM , Rating: 4
It wouldn't make financial sense for Apple to switch its consumer-side products to thin clients. Apple has historically made its money by selling consumers overpriced hardware that performed well. If they switched it over so that they sold thin clients, that means that those terminals could last a very long time (since the server is the computer doing all the processing) and consumers wouldn't need to buy the "next big thing" every year as they currently do.

At my old job, we used Winterms (thin clients) for years and years. There wasn't any reason to upgrade the terminals because all the changes happened at the server side. Apple does not want their hardware to be commoditised like that.


RE: Very ambitious.
By Jeffk464 on 11/6/2012 8:47:39 PM , Rating: 2
I'm pretty sure google already has a leg up on the dumb terminal concept. The new chromebook is based on ARM and is already a slick working cloud device.


RE: Very ambitious.
By ritualm on 11/6/2012 1:50:12 PM , Rating: 2
Political hot-air and maneuvering.

Apple can pull it off because it can always blame its users for its own products' shortcomings:

iPhone 4 deathgrip = You're holding it wrong!
iOS 6 maps = You're navigating it wrong!
iPhone 5 color shift = You're taking it wrong!
Retina MBP image retention = You're using it wrong!
Cellular data overages because iOS 6/iPhone 5 uses LTE even when WiFi is forced on = You're WiFi-ing it wrong!

Coupled with Apple's desire to make more money off its sheepish followers and fanboys, of course they can!


RE: Very ambitious.
By lightfoot on 11/6/2012 1:55:31 PM , Rating: 5
It also allows Apple to control the platform even more. No more of this "installing Windows in order to run applications purchased outside of iTunes" nonsense.

Apple must have at least 30% of the revenue from every piece of software running on its hardware! Don't forget, Apple owns the hardware, they only let you use it.


RE: Very ambitious.
By foolsgambit11 on 11/6/2012 8:13:51 PM , Rating: 1
Don't know if you noticed, but Windows runs on ARM these days.


RE: Very ambitious.
By cfaalm on 11/7/2012 12:57:02 PM , Rating: 2
Yes, RT does. You still cannot buy Windows RT anywhere, though.


RE: Very ambitious.
By tayb on 11/6/12, Rating: -1
RE: Very ambitious.
By ATrigo on 11/6/2012 3:43:19 PM , Rating: 4
quote:
The camera "problem" on the iPhone 5 isn't actually a problem in the sense that almost every "point and shoot" camera has this problem.


That's exactly what Apple wants people to think. Granted, all small "point and shoot" cameras will face light induced artifacts at some point... the difference is that the iphone 5 is overly sensitive to it where the iphone4 and other cameras are not. Apple simply made a bad design trade off.


RE: Very ambitious.
By amanojaku on 11/6/2012 3:52:00 PM , Rating: 5
quote:
Apple issued free bumper cases to iPhone 4 users before fixing the problem.
After weeks of denial and criticism.
quote:
Apple wrote an apology letter about iOS 6 maps
After days of silence and criticism.
quote:
The camera "problem" on the iPhone 5 isn't actually a problem in the sense that almost every "point and shoot" camera has this problem.
Color tint is not a problem that all point and shoot cameras have, or even most.
quote:
The other two aren't widespread problems and they've both been fixed.
Says you. Apple never releases numbers, and goes as far as to cover things up. The retina display issue was a worldwide problem for at least three months (June-August), so it was pretty big. As for WIFI, there were at least three issues with it (WIFI not working, cellular instead of WIFI, and lack of AES) and it affected enough people to be mentioned in every technical journal and web site. For a company that claims its closed approach means perfection, it sure ain't perfect.


RE: Very ambitious.
By tayb on 11/6/12, Rating: -1
RE: Very ambitious.
By someguy123 on 11/6/2012 6:05:57 PM , Rating: 2
I'm not sure consumer reports is biased but that article is not very convincing. They don't show any comparisons except the oddly purple iphone 4 image after claiming to have tested on a variety of devices.

iphone 5's problem is caused by the glass itself, regardless of overly bright light sources. certain angles just exaggerate the tint.

http://img.svbtle.com/dcurtis_24268809838680_raw.j...


RE: Very ambitious.
By HoosierEngineer5 on 11/6/2012 4:30:59 PM , Rating: 2
"Coupled with Apple's desire to make more money off its sheepish followers and fanboys, of course they can! "

Quod erat demonstrandum.


RE: Very ambitious.
By NellyFromMA on 11/6/2012 2:38:54 PM , Rating: 2
Lol, MS did the exact same thing as a test by developing Win RT and experiementing with the desktop interface on it.

At first they thought simply recompiling for ARM would be great. They sure found out its not quite as simple as that though.

I applaud Apple for trying, but mostly because they are playing catchup on something thats already been tried and doesn't really work (for the purpose of replacing x86-64).

Very un-Apple. Good luck with that.


RE: Very ambitious.
By DanNeely on 11/6/2012 3:10:15 PM , Rating: 2
The problem is that (excluding .net apps) every 3rd party developer would also have to recompile for ARM as well (to include every 3rd party binary library they used); without that happening you'd have no clear way for consumers to know what apps would/not work on WOA. The result of that would (like every other compatibility issue windows has ever had) be blamed entirely on MS and be used to argue the Win8 was crap. Drawing a strict line in the sand is much safer.

Unless ARM gets to the point of being able to emulate x86 at near atom performance levels it's not going to happen.


RE: Very ambitious.
By Jeffk464 on 11/6/2012 8:49:57 PM , Rating: 2
Intel should be getting nervous if this trend continues they will be relegated to gaming PC's and high end business/government computing needs.


RE: Very ambitious.
By inperfectdarkness on 11/7/2012 3:42:18 AM , Rating: 2
It's also increcibly stupid. Apple benefitted most from switching to Intel architecture, which greatily benefitted Apple's interoperability with windows-based hardware (though not necessarily by direct action on Apple's part).

Switching to a different architecture framework--while the rest of the world chugs along on the stuff that makes windows work--will seriously hurt Apple; just like what happened in the 90's.

Can Apple do it? Sure. Should they do it? Not logically--but when does logic apply to Apple?


Apple ditching the professionals?
By Operandi on 11/6/2012 1:58:50 PM , Rating: 2
I have to wonder if this is just another sign of Apple ditching the content creation pro market in favor of the consumer content consumption market, iPhones, iPads. We’ve seen this with most recent version of Final Cut, and the lack of new Power Macs for quite some time.
If this is the trend Apple is following then it does make sense as they hardly need the power of Intel’s and AMD’s latest CPUs and ARM would do just fine. If however they want to keep playing in the pro and enthusiast market I’m not sure if ARM will ever match x86 in terms of raw performance.




RE: Apple ditching the professionals?
By ClownPuncher on 11/6/2012 2:16:57 PM , Rating: 3
Other than music, none of us have needed to do content creation on a Mac in years.


RE: Apple ditching the professionals?
By 91TTZ on 11/6/2012 4:34:59 PM , Rating: 2
You've never edited an image in Photoshop (or any other image processing software) or edited/encoded a movie you filmed?


RE: Apple ditching the professionals?
By croc on 11/7/2012 2:26:08 AM , Rating: 3
Does Photoshop even port for Mac's these days?


By retrospooty on 11/7/2012 9:24:54 AM , Rating: 2
"Does Photoshop even port for Mac's these days?"

What is a Mac ?


By bobsmith1492 on 11/7/2012 7:07:18 AM , Rating: 3
Photoshop has been primarily PC-based for years now.


By ClownPuncher on 11/7/2012 2:47:07 PM , Rating: 2
Funny. I used to do design work on a G5, because that's how long ago Mac's had any sort of advantage there. Now, with a much more powerful PC, I can get those tasks done much more efficiently.


RE: Apple ditching the professionals?
By CaedenV on 11/6/2012 3:26:26 PM , Rating: 3
That was my first thought when I first heard this rumor a few months ago... but then I started looking around. Creative departments in big companies (like P&G) are largely moved to PC, both of the local TV stations in town (local cable) have moved entirely to PC, nearly all of my friends I went to school with (went to school for audio and video post production) are using PC. The only people I know using mac are students, because it is required, and most of them have installed win7 on their macs because the adobe suite runs better on the PC side, and some of the old school veterans who will not likely be moving platforms any time soon because it would be painfully expensive to do so. But still, content creation is not the same world it was 10 years ago... or even 2 years ago. They are bleeding market share very quickly on this one front.

Apple gave up on content creators, or at least they are unsure of what to do with them. I think their devices will always be suficient for fun little stuff (like loading ready run, and other fun stuff like that), but in the end I think they are just going to say 'screw it' in the end and give up on the desktop market eventually.


By chemist1 on 11/6/2012 11:08:09 PM , Rating: 2
There is one additional group among whom Macs are very popular -- scientists, particularly physicists and chemists. [This, incidentally, counters the frequently-repeated myth that people just buy Macs for style. Hell, we generally don't care about style -- just look at how we dress!]

Based on my own experience, I think there are four key reasons for the popularity of Macs among scientists: (1) OSX offers a native Unix environment for development work; (2) OSX is significantly more powerful than Windows for very complex office productivity tasks (say, where you have ten programs open, with ten windows in each one, and need to navigate rapidly among them -- I find this easy on a Mac, and hopeless in Windows); (3) broadly speaking, OSX offers a more intuitive interface--for scientists, our computer is just a tool, and whatever enables us to most easily and quickly get the job done is what is preferred (our work is challenging enough; we don't need the software to provide additional challenge); and (4) historically, much less bother about security -- with responsible use of either a Mac or a PC you're not going to get a virus; but responsible use of a PC has generally required installation and background operation of antivirus software which (at least in the past) has been a PITA, while responsible use of a Mac has not.

From the above, it should be clear that scientists do a lot of content creation on their computers. Given this, I share the concern that Apple seems to be moving away from content creation and towards content consumption, since I think the former has been its great strength.


RE: Apple ditching the professionals?
By TakinYourPoints on 11/6/2012 3:43:02 PM , Rating: 2
Their focus away from first party support of professionals has been pretty steady over the last few years. Professionals made up a very large part of their income and userbase when I started using Macs about ten years ago. Now they are mainly servicing the mainstream, which is fine (and clearly way more profitable) except that they are devoting fewer and fewer resources to pros.

Third party support is obviously still there, but things like what they did to FCP and slower updates with their workstations makes it very clear where Apple's priorities currently are.

Either way, the rumor of them moving away from Intel comes up every few months. The transition from x86 to ARM also seems like it would be much more difficult than the one from PPC to x86, so we'll see if it even happens. It does seems believable on a very long timeframe given that efficiency is such a high priority for Apple, and the fact that they are now designing their own mobile SoCs. It'll all come down to if it is worth it. If there's one thing Intel has proven themselves excellent at, it is having one of the best tech roadmaps on the desktop and laptop side.


By Master Kenobi (blog) on 11/6/2012 4:35:54 PM , Rating: 2
Agreed. Intel knows exactly what they are doing and honestly it was a good move on Apple's part to switch to them originally. I can understand them wanting to bring that piece in house, but the reality is they will never be able to outgun the biggest gun in town. Intel reigns supreme in the area of process manufacturing and their chip designs are second to none in the performance realm. If Apple wanted to bring that all in house, they need to get with TSMC or someone else to get their designs mass produced. Intel is typically a process node and a half ahead of anyone else in the industry at any given time and they devote a huge chunk of money into R&D to maintain that.

I think the biggest problem for Apple at that point would be phones and tablets using Intel chips and either Windows or Android and in performance benchmarks getting completely spanked because their chips just can't handle it.


Performance difference?
By edge929 on 11/6/2012 1:59:41 PM , Rating: 2
I don't see how ARM chips can match the much more mature x86 chips in raw performance. Sure, ARM chips are more efficient per watt, that's how they were designed but at the end of the day, I'll take a 50% jump in performance for 60% more power usage.




RE: Performance difference?
By FITCamaro on 11/6/2012 2:27:47 PM , Rating: 2
Exactly. They'll maybe keep the sheeple who buy their crap regardless of its performance, but for the 90% of people I know with Macs who do development, they'd never buy another one again. ARM doesn't hold a candle to x86 in performance.


RE: Performance difference?
By TakinYourPoints on 11/6/2012 3:45:49 PM , Rating: 2
Yup. Developers love having a Unix platform that is well supported with lots of applications, not to mention one that runs Windows well in a VM. Take that away and they'll have to use something else (Ubuntu perhaps?)


RE: Performance difference?
By Mint on 11/6/2012 6:44:18 PM , Rating: 2
Even per watt, they're only more efficient because Intel didn't bother going after the market. Atom was a half-assed effort good enough to prevent AMD/Via/ARM from getting the netbook market, but bad enough to make the ULV Core CPUs as desirable as possible.

The tablet/smartphone market had too low margins for their taste at first, but now it's big enough and premium enough to warrant attention. Clovertrail already virtually closes the perf/W gap, and Intel hasn't even put its best foot forward either in architecture or process. Just wait until they go 14nm.


RE: Performance difference?
By Jeffk464 on 11/6/2012 9:02:40 PM , Rating: 2
So if you have an x86 android phone are you going to have the same problem you have with windows RT where existing programs wont run?


RE: Performance difference?
By someguy123 on 11/6/2012 9:58:57 PM , Rating: 2
They can run dalvik apps but not NDK.


RE: Performance difference?
By Jeffk464 on 11/6/2012 10:01:51 PM , Rating: 2
argh, this bouncing back and forth between chips is going to be a pain for consumers.


RE: Performance difference?
By Jeffk464 on 11/6/2012 8:58:56 PM , Rating: 2
Actually thats not a good trade off if you are working off a battery.


Pretty ironic
By tayb on 11/6/2012 3:32:41 PM , Rating: 2
Merge the two product lines? But why? Just last year Apple talked about how crazy it was to try and merge a fridge and a toaster!

Honestly though it seems like it would make more sense to wait for Intel to catch up and pass ARM than it would to switch OSX to ARM.




RE: Pretty ironic
By Camikazi on 11/6/2012 4:49:10 PM , Rating: 3
Steve Jobs is gone and now Apple is doing a lot of things it said it wouldn't before.


If it happens
By Shadowself on 11/6/2012 2:56:52 PM , Rating: 2
IF it happens, it won't happen before 2020 at the earliest.

The MacPro line won't move over to ARM before then for sure. Apple is just as likely to kill the MacPro line as to move it to ARM.

The iMac line is possible by 2020, but very unlikely before then. Apple is still shipping i5 and i7 chips in the iMac and likely will be shipping the equivalents through 2020. The processing power in even the newly announced (and details in rumors) for the 64 bit ARM designs won't catch up to where Haswell will be until at least 2020.

No matter what there will be a step back in performance in going from the then i7 or Xeon equivalent of whenever the switch is made and the then current ARM system. ARM is a reasonable compromise between performance and power draw. The i7 and Xeon chips weight the equation very heavily in favor of performance and NOT in favor of low power draw.

Also note that the focus of Haswell and Broadwell are on the lower power end (not as low as ARM, but certainly lower than Intel has focused on before) AND that the Intel has yet to even informally announce the dates for the Xeon versions of Ivy Bridge.

Intel knows that going forward their profits are going to come from lower power chips. In the next five years Intel is going to sell many, many more "mobile" chips (think i3 and i5 equivalents) than workstation ships (think i7 and Xeon). Intel is going to follow the money.

Some people have gone so far as to say that Apple's chip design teams are larger than the teams at DEC or SUN when they started their own workstation CPU design efforts -- as if that is justification enough for Apple to design its own desktop related chips. And what happened to Alpha and SPARC?

I don't expect Apple to make its third chip switch (1. 68k to PPC, 2. PPC to x86 and 3. x86 to Ax) for the Mac before 2020. I'd be truly shocked if it happens before 2017.

The last fly in the ointment is the Mac's Windows compatibility. While some may claim it is just coincidence that Mac sales have increased since Apple started directly supporting people Windows on the Mac. Apple must know that Windows support is critical to the Mac. I doubt Apple will ever get Microsoft to tweak Windows for an Apple specific implementation of the ARM designs.




RE: If it happens
By DanNeely on 11/6/2012 3:12:15 PM , Rating: 2
quote:
Some people have gone so far as to say that Apple's chip design teams are larger than the teams at DEC or SUN when they started their own workstation CPU design efforts -- as if that is justification enough for Apple to design its own desktop related chips. And what happened to Alpha and SPARC?


I'd certainly hope they are. A modern SOC dwarfs an 80s/90s era RISC chip in complexity.


lol, rumors!
By MadMan007 on 11/6/2012 1:44:18 PM , Rating: 3
Report: Analysts need to pump up their Apple holdings, make up rumor.




Good news
By Ammohunt on 11/6/2012 1:52:27 PM , Rating: 2
I am all for ARM based desktop platforms just can't afford the Apple premium right now. Hopefully their will be an open desktop standard in the future for ARM systems.




Ha? BS!
By yvizel on 11/6/2012 4:04:06 PM , Rating: 2
Yes, I would believe this when I see it.
And yes, it is better to wait for Intel to pass ARM. I would bet on the next Atom architecture to pass ARM.

Another thing, just look at Ivy Bridge. A15 is no where NEAR this kind of performance.

So if this story is true, then people believe that Apple (a software company) can come up with an ARM design that is 10X faster than its current A6 in two years. Than, two years after it comes with another 10X improvement.

Yeah right...




By drycrust3 on 11/7/2012 1:49:08 PM , Rating: 2
quote:
On the other hand, ARM is invading the personal computer market with Windows RT, and is expected to land in the server market around 2014 with Window Server 2012 R2.

I did a search on the DistroWatch.com website for "ARM" and "server" and got 8 results: Fedora, Slackware, FreeBSD, PUIAS Linux, etc, already have a server distribution that runs on ARM. Here is the link for CRUX: http://distrowatch.com/table.php?distribution=crux
Also, today openSUSE say the have released a desktop version for the ARM processor: http://distrowatch.com/?newsid=07539




Thinking big
By Tony Swash on 11/6/12, Rating: 0
"We’re Apple. We don’t wear suits. We don’t even own suits." -- Apple CEO Steve Jobs














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki