back to article Four-pronged ARM-based Mac rumor channels Rasputin

An intriguing rumor is a difficult thing to kill – and the long-whispered switch by Apple of its Mac line from Intel to ARM-based chippery has again resurfaced, Rasputin-like, to live another day. Jeffries & Co. industry analyst Hyunwoo Doh, Barron's writes, has released a report that recommends his clients stock up on Samsung …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Anonymous Coward

    [Citation needed]

    Well, not a citation so much as a healthy dose of scepticism about anything an analyst says. In particular

    - who would benefit from this news being published now?

    - what is the relationship between the author, their employer, and any of the companies mentioned in this report?

    - what is the analyst's track record of being right?

    - what is the penalty for being 100% wrong?

    An analyst can spout nonsense knowing that it doesn't really matter if they get it utterly wrong (no one at Jefferies is going to get fired if Apple decide to stick with Intel, say). We also don''t know whether Doh or Jefferies has a relationship with (say) Samsung that would be enhanced by the latter stock rising. After all, this is not an isolated report by a technical site like Tom's Hardware. The claim we're reading about here was made as part of a larger report that rated Samsung stock as a buy.

    #1 rule in journalism: reporting is not the repetition of hearsay.

    1. Tim99 Silver badge
      Coat

      Re: [Citation needed]

      I thought the #1 rule in journalism was to print as many people's names as possible in the article, and spell them correctly? That way all of them, and their friends, obtain the publication.

  2. John Savard

    Interesting Idea

    It was very hard for me to take this idea seriously at first, but it is true that there are apparent reasons why it would make sense for Apple to do this.

    However, there are no compelling reasons for not 'being at the mercy' of Intel's development schedule, simply because Intel's business is making chips for desktop and laptop computers, and so plenty of choice is available. It's not like Apple's reasons for changing from the PowerPC to Intel.

    So changing chips and thus condemning the Macintosh to perpetual inferiority to the Wintel platform doesn't make sense.

    Now, Apple could make a Macintosh II in a UMPC-type form factor based on ARM, and leave the regular Mac alone. This would be an attractive product to some, and it would fill an empty niche. It could be a very profitable product. But it somehow seems like a very un-Apple-like thing to do.

    1. John Brown (no body) Silver badge
      Joke

      Re: Interesting Idea

      iSurface?

  3. Anonymous Coward
    Anonymous Coward

    NAS/TV/Home Control box

    It would make sense for some that was left on 24/7, like a NAS, or TV or Home controller box - and there have been rumours of at least the last two of these recently.

    Imagine a box you could plug overpriced apple hard disks into, that would serve up content like an apple TV box, and that could integrate with home control boxes (also apple). Further, imagine you could VNC into Mac sessions running on said box from your ipad, giving you a tablet Mac at home or on the go (although it would have to be more proprietary than VNC, this is apple).

    Doesn't sound so dumb now.

    1. P. Lee

      Re: NAS/TV/Home Control box

      That would be a spec'd up AppleTV box or or likely, AirPort.

      Given that most of a tablet is the battery and screen, it would make more sense to add ARM chippery to x86 hosts so you can run ipad apps while the intel chip sleeps. That would open up the market for appstore apps in the desktop arena. Trialing then driving productivity apps with laptops running faster/less efficient ARM cpu's which will be refined and later trickle down to the real tablets and phones.

      I can't see Apple giving up x86 on a mac - there are plenty of buyers who need windows to run and their margins are not to be sniffed at. However, I suspect Apple would like to drive app development for ipads harder allowing them to push ipads as desktops, with even higher margins.

      ARM-based Apple home servers, hidden in AirPorts, would be the low-hanging fruit. Fast local sync and then slower replication over the WAN to the cloud allows more complete replication to take place.

      I wonder if anyone else might be doing local server work with replication out to the cloud... oh, hello Valve!

      1. Mage Silver badge

        Re:Giving up x86

        I think out side IT circles, i.e. most of Mac user base, running Windows or Linux on Mac is very very rare. It's anecdotal, but none of the Arty types I know have ever run other than Apple's OS X. Or in the case of very long term users, previously OS 9.

  4. Bruce Hoult

    I've got an iPad with an A7 in. It benchmarks and feels nearly identical to my 2.5 GHz core2duo laptop. That's plenty for a MacBook Air style laptop even today.

    1. Salts

      Or a Chromebook type Mac, but I don't see the Macbook Pro going ARM anytime soon, I have a great deal of respect for ARM and like to see them do well, but I doubt they can match a quad core i7, well not just yet...

      1. Zolko Silver badge

        Transformer

        If you add a keyboard to an iPad, you get the best of both worlds. But of course, Asus has done that before, therefore Apple would be hard pressed to pretend to innovate.

  5. Daniel B.
    Happy

    ooooh

    I'd love to see a non-x86 Mac in the near future, but I'd also like to see them perform better than the craptel stuff as well. It's about time someone brings back RISC on mainstream computers...

  6. Anonymous Coward
    Anonymous Coward

    Analyst = STUPID

    If Apple switches to ARM for Mac laptops, they can only convert at most half of them or so, because half of Mac laptop buyers want to run Windows apps at least once in a while. That limits the market to a few million a quarter, or a less than 10% increase over the number of ARM SoCs Apple is using in iPhone and iPad.

    And that, of course, assumes that Samsung remains the manufacturer of 100% of Apple ARM SoCs, when rumors indicate that they're switching to TSMC for at least a good chunk of them, and could easily switch to GF for the remainder if they wanted to remain Samsung-free (I'm not sure they do, but if they did...)

    The idea that one should buy stock in a multinational corporation like Samsung, of which their chip foundry accounts for a single digit percentage of their earnings (even if Apple accounts for the large majority of that single digit percentage) on the basis that Apple might increase the number of CPUs they order by a single digit percentage is silly.

    Even if Apple had Samsung continue to fabricate 100% of their ARM SoCs, and Apple was somehow able to convert every Mac to ARM, it would mean an increase in Samsung's revenue by maybe 2% at best. Yeah, I want to buy stock to capture that! I would think a better play would be shorting Intel, since any switch of Apple from x86 to ARM is going to have a larger downward impact on Intel's share price than an upward impact on Samsung's share price, since Intel is almost exclusively a chip provider, while Samsung has businesses ranging from toasters to oil tankers of which their foundry business is only a part.

    1. Goat Jam

      Re: Analyst = STUPID

      "because half of Mac laptop buyers want to run Windows apps at least once in a while"

      Well, that may or may not be true, but it is also true that MS are currently working really hard to merge all their Windows lines to a point where it doesn't matter what type of Windows you have your code will run on it.

      What is the figure MS put out a week or two back? I think they said that 80% of code can be cross platform compatible right now. Of course that does not mean that things are currently quite like that in reality, but if MS are successful in that endeavor then there is no reason that an ARM sporting Mac couldn't run Windows RT (or its successor) instead.

      Personally I doubt that MS will succeed with their One Ring strategy but stranger things have happened.

      1. Anonymous Coward
        Anonymous Coward

        @Goat Jam

        I doubt Windows RT will still be around in a couple years. It has been an abject failure, what little success Surface can claim has been from the Pro model that runs x86.

      2. GreenOgre

        Re: Analyst = STUPID

        > "because half of Mac laptop buyers want to run Windows apps at least once in a while"

        So? Push them up the margin curve to the MacBook Pro. If they NEED S+MS Windows, they will willingly pay through the nose

    2. Anonymous Coward
      Anonymous Coward

      Re: Analyst = STUPID

      "That limits the market to a few million a quarter"

      You're forgetting what *new* capabilities this device might have. An OSX laptop with a three-day battery life perhaps? Which also runs iPad apps? With built-in LTE and phone calling capabilities? It could open up completely new markets.

  7. A Non e-mouse Silver badge

    Hmmm....

    At the lower end, I could see a MacBook Air with an ARM CPU being an interesting idea. But ARM isn't know for it's higher performing CPUs, so the higher end MacBook Pros and all the desktop Macs I can't see moving to ARM. As others have mentioned, by having an ARM CPU machine, you would also loose the ability to run Windows on the machine.

    Also, running your O/S on multiple CPU architectures, and persuading your ISVs to write their apps for both CPU architectures doesn't doesn't sound easy. I know iOS and MacOS are based on the same kernel, but there's more to an O/S than just a similar kernel (Just ask all the Linux distributions!)

    IMHO, the best way to make this fly would be to get the Transmeta people they took over to crank out ARM/x86 translation tools.

    1. GreenOgre

      Re: Hmmm....

      The difficulties between iOS and MacOS are in the screen size and input methods.

      For most Linux apps I have tried to compile for ARM it has simply been a case of changing an "available architectures" flag.

      For applications running on Unix-like OSes such as iOS, OSX and Linux, changing architectures is almost trivial once the compilers are available for that arch.

      However, there is still the Human Interface element of making your windows and dialogue boxes usable on both an iPhone and Thunderbolt display! But then, that's why you developers get paid the big bucks! :-)

  8. Anonymous Coward
    Anonymous Coward

    Doh is a dear

    ...for sharing that.

    "According to Doh (no Simpsons jokes, please – we're adults here...),"

    1. MrT

      Re: Doh is a dear

      Yes indeed - even avoiding the Simpsons, the comedy potential has a long, long way to run...

    2. John Brown (no body) Silver badge

      Re: Doh is a dear

      "According to Doh (no Simpsons jokes, please – we're adults here...),"

      I was just wondering when the Reg demographic had changed so drastically and why were we not told.

  9. Anonymous Coward
    Anonymous Coward

    Difficulties and Implications

    If they do switch it's going to be very difficult for Apple to carry Mac OS X over from Intel to ARM.

    When they've previously switched architectures they have always had the benefit of moving to a faster CPU. This has made the emulation layers necessary for supporting old binaries perform acceptably fast for the end user.

    However, good though ARMs are they're not Intel i7 fast. Emulating x86/x64 code on ARM is going to be horribly slow. One off conversion of binaries from x86 to ARM might result in a program that works OK but it's quite often not permitted by EULAs.

    So Apple would either have to somehow persuade everyone to recompile and retest their code (a mamouth task with the added imperative of there being no viable alternative), or ditch the whole OS X ecosystem altogether. So that might mean an ARM MAC is an iOS machine instead.

    1. GreenOgre

      Re: Difficulties and Implications

      Who cares if it's not as fast as xyz processor?

      My car isn't as fast as an Aston Martin but can carry Wife, kids and Mother-In-Law across town in a third the time it would take in the Aston. (Hint: the Aston would need two trips)

      I don't mind carrying an ARM Chromebook in my carry-on, I prefer to leave my quad-core i7 "laptop" on a desk that's braced to support it.

      ARM is plenty fast enough for 90% of day-to-day tasks, though perhaps that i7 laptop can save you a bundle on your gym membership?

  10. hungee
    Facepalm

    ha!

    Quick check here...

    Baytrail 3740 atom x4core benchmarks similar synthetic CPU scores to the A7 x2core in the iPad Air

    Now let's look at some benchmarks for x86

    PCMARK SCORES

    Baytrail ~1,200

    i7 laptop ~7,000

    Top Xeon desktop ~17,000

    As if Apple would be so stupid as to ruin performance so massively by between 5-15x

    The code would have to be super-efficient compared to x86 to even get close.

    A macbook air? Maybe. But then you have to deal with 3rd party application compatibility issue and that is not simple to explain.

    1. Dan 55 Silver badge

      Re: ha!

      It wasn't that difficult last time sound and is even easier now with an App Store. People could live with it if a MacBook Air had e.g. 20 hours battery life.

    2. Anonymous Coward
      Anonymous Coward

      Re: ha!

      "As if Apple would be so stupid as to ruin performance so massively by between 5-15x"

      Unless they plan to put say 16 or 32 ARM cores in a chip, and power them up dynamically? And/or add their own instructions to optimise emulation of Intel code?

  11. Mage Silver badge

    Maybe it's not true now.

    Perhaps this is some other product, a large iPad with keyboard cover?

    But eventually Apple will leave niche "high performance" users in the lurch. They will aim for volume market, iTunes lock in etc. So eventually the ARM will replace x86 for Apple. Unlike MS, they switched before. 68000 and then Power PC. So they have no loyalty to a CPU ecosystem and certainly the possible deprivation of rebooting to windows will be low priority.

    Nothing new, it's been said before. I've said it before. But I think 2015 - 2016 more likely.

  12. Torben Mogensen

    Reason #5: Control

    Apple has (almost) always wanted as much as possible control over their hardware platform, both to tune hardware and software to each other but also, equally important, to prevent people running their software on machines not made by Apple.

    By making their own System-on-Chip (SoC), they can pretty much ensure that only computers made by Apple can run their software. As for performance, the gap between ARM and Intel is dwindling now that ARM has a line of 64-bit processors. And, as the article said, MacBooks are not known as power houses anyway.

    Somebody mentioned migration as an issue, but this is much less an issue than it was at the previous processor switches (from 68K to PPC and from PPC to x86), as most software is now written in high-level languages that can easily be compiled for other processors. Also, MacOSX and IOS themselves already share a large code base that is just compiled for the different platforms.

    So, technically, I see no major hindrance. The main reason to stay with x86 would be if Apple could pressure Intel to give them really good prices by threatening to switch if they don't.

  13. chivo243 Silver badge
    Boffin

    As I posted before in another thread

    "Think harmonic convergence here man! Apple has a road map to unify OSX and iOS, that was Steve’s plan for global domination..." 4 down votes too! Care to think again?

    An ARM CPU in a Mac would make this so feasible.

  14. Norman Hartnell

    "we're adults here..."

    Umm...sure.

  15. Tachikoma

    Does this mean I can run iTunes on my Archimedes?

    1. Wilseus

      Does this mean I can run iTunes on my Archimedes?

      Why would you want to? Ugh!

  16. Wilseus

    ARM vs. x86

    This is one of the most interesting topics I have seen for quite some time. I remember Acorn's glory days when the ARM beat the pants off the Intel chips of the day, which is what the original chip was designed to do.

    Common sense would dictate that any chip supporting any sort of RISC instruction set, containing the same number of trannies (put to good use) as a modern x86 chip would thoroughly and convincingly trash the x86, because it has to dedicate so much silicon to sorting out the hideous mess that is the x86 instruction set which anyone who has ever dabbled with assembly code will know.

    But is this true? Why did so many promising RISC designs that had such high performance (SPARC, Alpha, MIPS, Power) fall by the wayside? Was it simply market forces, or is the efficiency of RISC systems just not that important, as Intel keeps saying?

    Before anyone says it, sticking 32 or 64 ARM cores on one chip might work for servers but won't work for general desktop use, see Amdahl's Law

    Why are there still no RISC systems that thoroughly outperform x86 on a core for core basis?

    I'd love to see ARM (or MIPS) based laptops and desktops out there beating the socks off Intel's highly-polished turds, but there must be some reason that this has not already happened. I just don't really understand what that reason is.

    1. Wilseus

      Re: ARM vs. x86

      Just to add to what I said earlier - and apologies for taking this a little off-topic - but this link will help illustrate just how much of a cluster**** x86 really is. Be warned, it's not pretty!

      In contrast, you can summarise the entire instruction format of a typical RISC processor in just one page.

    2. GreenOgre

      Re: ARM vs. x86

      > I just don't really understand what that reason is.

      "It's not what we're used too"

      Most IT folks and "power users" are still mired in the limited and restrictive Wintel "ecosystem".

      Viewing the comments above shows that even fairly knowledgeable people are entirely unaware of how easy it is to port code from a modern operating system between architectures. Binaries for ARM systems can be compiled on Intel platforms (or vice versa) alongside their native versions FROM IDENTICAL CODE.

      An OS vendor (e.g. Apple) simply has to provide the tools and a couple extra check boxes in their SDK.

      Build package for ?

      1: MacBook Pro

      2: MacBook Air

      3. iPad

      4. MacBook Vacuum

      5. All of the above

      Upload to Apple Store? Y/n

      Job done!

    3. FutureShock999

      Re: ARM vs. x86

      "Why are there still no RISC systems that thoroughly outperform x86 on a core for core basis?"

      There are a number of reasons. The first one is that Intel has a huge investment in fabrication plants ("fabs" in industry parlance), and usually manages to be a generation ahead of the ARM builders in implementation technology - lithography, insulation, and/or feature size. They just have a better _chip_, regardless of the architecture. Samsung and TMSC are catching up, but for a long time Intel has a huge advantage simply in construction - and this gave them faster clock rates, lower voltages, etc.

      The second reason is that convoluted x86 architecture makes no claims to be elegant or simple. But it IS well known what the asymmetries are, and Intel's compilers are heavily optimised to work with them where possible. So having finely tuned compilers has greatly helped even-out the architectural differences.

      Thirdly, Intel simply invests a whole lot more into developing their x86 follow-ons than ARM has. Such is the result of market dominance. Intel really does have a number of very smart designers, and as even you stated, it may be a turd, but it is a WELL-POLISHED turd. Things like branch-prediction, out-of-order execution, superpipelining...Intel has these down to a very fine art, with a degree of circuitry dedicated to this that ARM historically lacks.

      Lastly, for ARM to focus on execution speed, it would need to add lots of transistors on-die to do these things (Out-of-order, prediction, superpiplining) to the degree that Intel does - and THAT would negate a good portion of ARM's power consumption advantage. There is no free lunch - having a better instruction set doesn't make up for lacking the degree of dedicated silicon that Intel uses to speed up execution. The day that ARM begins to catch Intel in single-core execution speed will also be the day that ARM begins to catch Intel in power usage...

    4. Dig

      Re: ARM vs. x86

      Simple explanation. Every op code in arm will have a direct or very similar mapping to an op code in x86 so already both will be the same performance, excepting process differences, any extra will then come down to memory bandwidth utilisation. This is where the cisc architecture then kicks in as a single opcode(OK it may be multiple words) is then converted to multiple intermediate instructions but using less bandwidth on the bus.

  17. Mike 16

    Misc Musings

    First off, the silliness. If you think an iPad is a fully capable replacement for a MacBook, you must not edit many Makefiles. Paying extra for a TAB key just enrages some folks.

    Some history. A friend worked at Apple back in the day, and his group produced a IIGS followon that was ARM based. Ran all existing (6502-based) IIGS code. Snappier GUI than the then-current Macs, cheaper, oops! So it was "gassed".

    If you-all think that "just re-compile" is so easy, and Apple so supremely competent at re-targetting their software to new platforms, perhaps you can explain why their special flavor of X was so badly broken by the transition to x86? This was software that had run either (and even "cross") endian for over a decade and they managed to introduce rookie endian-bugs. Not to mention that even when they went from 68000 to 68020 they managed to stumble over the "let's just stick some unrelated flags in the upper byte of these pointers" bug that had bedeviled the 360->370 transition, again, a decade before.

    Not to say it won't happen. They may be able to hire someone less Laurel-and-Hardy to do software (for a change). And the move to their own ARM SOC would indeed be a master-stroke for "you will get all your software via iTunes/App-store, and will update when we punch the button, and will not whimper or your device will die", which is so clearly the path forward.

    1. Wilseus

      Re: Misc Musings

      > A friend worked at Apple back in the day, and his group produced a IIGS followon that was ARM based. Ran all existing (6502-based) IIGS code. Snappier GUI than the then-current Macs, cheaper, oops! So it was "gassed".

      That doesn't surprise me one bit. Their early familiarity with ARM also makes it even more interesting that after the Power Mac was released, their marketing department "forgot" that it wasn't the first RISC home computer.

      > If you-all think that "just re-compile" is so easy

      I certainly didn't think that! In fact I am working on this sort of thing right now, porting a large amount of x86/ARM targeted C code to another CPU. Setting up a cross-compile environment is far from easy, even if the tools you are using actually support your target CPU properly (In my experience, "experimental" invariably means "unusable.") and it never "just works" even when you eventually get the damn thing building!

This topic is closed for new posts.

Other stories you might like