Here’s your Apple A9 and A9X confusion thread

We learned from today's announcements that Apple has produced a couple of new, high-end mobile SoCs for its iPad Pro and iPhone 6S products, and the company made some pretty serious claims about the performance of those devices. Let me outline for you what we know, and then it will be clear how much is missing.

First, the iPad Pro is powered by the A9X SoC. Apple says its CPU is as much as "1.8x faster" than the triple-core "Cyclone++" processor in the A8X chip onboard the iPad Air 2. That enhanced version of the Cyclone core was already one of the most potent mobile CPUs cores anywhere. The firm also claimed on stage that the A9X CPU is faster than 80% of the portable PCs shipped in the last 12 months, which seems like quite the assertion. GPU-wise, the improvements are apparently even larger: twice the performance of A8X and faster than 90% of the mobile PCs shipped in the past year.

That's not all. The A9X's memory bandwidth has purportedly doubled, as have its read/write speeds for flash storage.

The A9 chip that drives the new iPhones is similarly improved, with Apple claiming up to 70% faster CPU performance and and 90% faster graphics.

The A9 is said to be Apple's third-generation 64-bit processor, so it presumably has a further evolved CPU microarchitecture. I would expect the improved CPU core to process more instructions per clock cycle than prior generations, perhaps at higher clock speeds or with added energy efficiency.

The firm also revealed that the A9 chip is based on a "new transistor architecture," which is almost certainly FinFETs at 14 or 16 nm. The new process technology should allow for lower voltage operation, improving the SoC's power efficiency in ways that can be converted back into performance. That change alone could contribute a sizeable portion of the CPU performance gains.

Here's what we don't know: everything else.

The A8 SoC has two cores while the A8X has three. Has Apple maintained the same core counts in the A9 and A9X? Given the similar claims it's laying down about performance claims for the two SoCs, I would expect so, but the firm didn't say.

Critically, we also don't yet know how much RAM is in the new iPhones and iPads. The iPad Air 2 is the only iOS device of the the past generation with 2GB of RAM onboard, and I view 1GB of RAM as the single biggest weakness of the iPhone 6 and 6 Plus. I'd bet on 2GB in the new models, but I'm an optimist.

Then there's the bit where Apple says the A9X is faster than 80-90% of mobile PCs shipped in the past year. What is the basis for comparison there? I'm not talking about the CPU and GPU tests being used to measure, which are an utter mystery, but about the PC processors in question. Does Intel ship so many Atoms and Celerons that the A9X can eclipse such a large proportion of mobile PCs in performance, or is the A9X able to outperform some portion of the chips based on Broadwell and Haswell?

Apple has more money than God and may have spent a small slice of it building a world-class CPU architecture. But has it really caught up to Intel's big cores? Here's the thing: I dunno!

That's just the tip of the iceberg of ignorance we all, collectively, have about the new Apple SoCs. I expect some details will filter out today and in the coming weeks as the devices based on these chips get into the hands of the press and public. Feel free to post in the comments below as that process happens. Perhaps we'll learn something, or perhaps we'll just multiply our ignorance. Hard to say.

Comments closed
    • torquer
    • 4 years ago

    When you design the hardware and the software, you have an advantage no other manufacturers can claim. Samsung will never be able to tap into some of the efficiencies that Apple can simply because they don’t own the OS. Even as open source, you’re talking about a generic OS running on either generic or custom ARM cores.

    The best thing about Apple is they control the whole thing (other than the ARM IP itself). If you’re happy in their ecosystem, nothing else comes close.

    If you’re not happy living by their rules, well…

    • slate0
    • 4 years ago

    Is it too early to ask if there is h.265 hardware support?

    • kamikaziechameleon
    • 4 years ago

    Speaking specifically to the topic here, GPU wise Intel has been essentially loosing since Tegra started rolling out. their performance per watt has indicated Intel was finished in this arena forever ago. Intel has been over celebrated for their GPU offerings improving relative to intel standards vs relative to universal standards. But as tegra, apple, and arm have moved forward leaps and bounds in the past 3 years while intel has not really. I mean the ability to drive pixels on a mobile solution is way higher than on intel…

    I understand that there is some behind the scenes magic here but to the laymen intel lost this war 3-4 years ago when they felt their phone games ran better than their laptop games did.

    Simultaneously Intel has been fumbling their low power offerings for some time. If you’ve used a lower power offering from intel and compared that experience to a modern flagship phone or tablet experience and you realize how immediately how advanced the tablet/phone software and hardware marriage has become on the cpu side.

    • kamikaziechameleon
    • 4 years ago

    double post sorry.

    • kamikaziechameleon
    • 4 years ago

    Phone GPU’s have been eating Intels lunch for some time. This is nothing new. For a fraction of the cost and power they achieve near identical results. If phone/tablet nvidia/arm gpu’s meant to compete with Intel they’d win with little effort.

    That’s the joke of buying a more expensive macbook with inferior GPU capabilities… but for whatever reason the incompetence of intel GPU’s is something we don’t talk about anymore cause their graphics drivers work now. But they are clearly still not offering truely relevant hardware to us, and on that front they are getting tucked into premium products and still offering amazingly under-powered results.

    Side Note, the iWatch stuff is getting increasingly interesting. After my hands on time with one I’m sold soon as I got the loose change to grab a gen 2 offering.

    • SuperSpy
    • 4 years ago

    Adobe let slip (Ars reported on it, I think) that the iPad Pro’s A9X would have 4GB of RAM, so I would about guarantee all the A9 devices ship with 2GB.

      • September
      • 4 years ago

      Hopefully the regular A9 has 2GB and the A9X *might* have 4GB. It is after all a 64-bit processor and a “Pro” device. Also, why else double the memory bandwidth but to talk with more RAM on the board?

      Remember, one of the distinctive features of the “X” models have always been that the RAM is elsewhere on the circuit board and not POP on-top of the CPU like the A7, A8, A9. This has allowed higher clock speeds AND more RAM as it isn’t limited by the footprint of the SOC.

    • raghu78
    • 4 years ago

    Apple already had a CPU core in the Cyclone used in A7 which was comparable to Intel’s big cores like sandy/ivy/haswell. the only difference was it was designed for mobile and a freq range of 1.3 – 1.8 Ghz.

    [url<]http://www.anandtech.com/show/7910/apples-cyclone-microarchitecture-detailed[/url<] Apple was waiting for TSMC 16FF+ to unleash the full power of A9X and thats exactly what was happened. Apple is providing four full cores and not fooling around like Intel with 2 cores and 4 threads using SMT. [url<]http://www.anandtech.com/show/9615/apple-announces-the-ipad-mini-4[/url<] According to Apple A9X is 1.8x faster than A8X at CPU tasks, and 2x faster at GPU tasks. Given that A8X was already the fastest ARM SoC in a mobile device this is quite an accomplishment, and is likely the result of architectural improvements, higher clock speeds, and possibly the addition of a fourth CPU core. We'll have to wait until we get our hands on the iPad Pro before any of this can be confirmed though. Apple also noted that A9X is built on a new "transistor architecture" which means it's being fabricated on either Samsung's 14nm or TSMC's 16nm FinFet process. This chip is going to be a monster in CPU performance and will be much more faster than Skylake Core M 6Y75 [url<]http://browser.primatelabs.com/geekbench3/2637682[/url<] [url<]http://browser.primatelabs.com/geekbench3/search?utf8=%E2%9C%93&q=a8x[/url<] Skylake 6Y75 scores 5700 in multi core while the last gen A8X scores 4500 in multi core. A9X even with a conservative 60% improvement will be around 7200 and if it gets close to 80% increase as Apple said then A9X will cross 8000. Thats mind boggling for a 5w mobile SoC. Apple is going to force Intel to add cores to their Core M chips. If Kaby lake does not do it Cannonlake at 10nm will surely do it. Intel can no longer play the fool with providing the same 2 cores / 4 threads and charge USD 300 - 400 for the Core M. Apple is doing the role which AMD should be doing - Pushing Intel harder to give more performance instead of just protecting its record margins. I am looking forward to Apple bringing SMT in A10X and also what it will do at 10nm. Apple has been super aggressive in pushing mobile performance and Intel should learn from them instead of trying to protect their record margins.

      • Speedfriend
      • 4 years ago

      One thing I have realised about Apple is that management only tell you what they want you to here and you can draw more conclusions from what they don’t say. So I wonder what conclusion we can draw by the fact that A8X has not gone into the iPad Mini 4 and that we didn’t see an iPad Air 3 with the A9X in. Is the A8X’s TDP too high to go into the Mini 4? And similarly, is the A9X too high to go into a normal sized iPad? Perhaps it is above the 5W range.

      My guess is that Apple have used the move to 14/16nm to add an additional core. It is telling that iPhone 6s/6s+ battery lives haven’t changed and that screen resolution hasn’t gone up. So something about the new SOCs design is eating power, otherwise the move to 14/16 should have allowed battery lives to increase.

      Last time Apple said the A8 had 25% better CPU improvement in iPhone, but what was actually delivered in Geekbench was 15% (or if you compare the A8 at 1.4ghz to the A7 at 1.4ghz in iPad Air, then the improvement was only 9%). So it will be interesting to see if the 70% performance gains are in some specific benchmark and whether they are reflected in geekbench.

        • Beelzebubba9
        • 4 years ago

        I think cost reasons are a big driver as to why Apple didn’t put the A8X in the budget iPad. The A8X is just a large chip by any standards, and third core and 128 bit memory controller likely drive up the cost substantially relative to the A8 and that doesn’t make much business sense.

      • the
      • 4 years ago

      Core count has yet to be confirmed anywhere to my knowledge. The idea that Apple has increased clock speeds and with some architectural improvements can reach that 80% figure. Remember that ARMv8 is an entirely different ISA than x86 so there could still be plenty of efficiency gains compared to the refined Sky Lake chips.

      I also don’t see Intel adding more cores to be competitive with Apple. Why? Who else does Apple sell their chips to? No one.[super<]*[/super<] Intel and Apple are not direct competitors between the Core M and A9. It is true that the iPad is a competitor with other tablets/phones that may use x86 but that is a comparison at the [i<]product[/i<] level. Furthermore, where is the actual need for more cores compared to a faster dual core in iOS based products? Only recently has Apple really enabled multitasking so there could be an argument for that going to 3 or 4 cores could merit some benefits. However, how many applications that you'd actually multitask with continually utilize a full CPU core? Very few come to mind. Apple has wisely stay out of the race for increasing core count while focusing on individual core performance and system level optimizations (L3 cache!). I do see merit behind adding SMT (aka Hyperthreadng in Intel speak) to Apple's CPU designs. This would increase throughput with a highly threaded or multitasking scenario. More importantly, I see SMT has being more energy efficient compared to adding additional cores to accomplish the same thing. Power efficiency trumps pretty much every other design aspect in ultra mobile. *Actually Apple is still under contract to sell various PA-Semi chips to the US military. Those are PowerPC, not ARM.

    • DavidC1
    • 4 years ago

    It could be a genuine improvement.

    This is a case of a well-executed, well-funded company versus those that are losing their focus and playing catch up.

    Of course the latter is Intel and AMD. The latter is worse, but lets talk about the former. They’ve been trying to catch up in the mobile arena and increasingly focusing EVERYTHING towards their, but to their misery, has been quite behind even in performance.

    While they are doing that Apple catches and beats their baby Core M. Actually its not that impressive considering how bad Broadwell Core M was. Broadwell Core M should have performed like a Haswell U, but it performs more like Sandy Bridge chips.

    But this isn’t going to affect Intel in the short term, even medium. The reason is what’s called a “Windows Shield”. Impenetrable defense protected by the corrupt government believing in schemes like “too big to fail” disallowing any potential competitors to x86 Windows market from entering.

    • iatacs19
    • 4 years ago

    I think you are reading too much into Apple PR specs. 😉

    • hasseb64
    • 4 years ago

    “A9X is faster than 80-90% of mobile PCs shipped in the past year.”
    Faster?
    “MIPS faster” or just faster to work with doing “normal” office tasks?
    We all know that 80% of laptops sold are cheap&slow POS.

      • adisor19
      • 4 years ago

      I think that last part is what Tim was trying to say : most PCs sold whether desktops or laptops are crap since even Apple tablets are more powerful. It hurts because it’s true.

      Adi

        • hasseb64
        • 4 years ago

        True!

        • derFunkenstein
        • 4 years ago

        I wonder if 80-90% of all portable PCs sold are running a combination of ARM, AMD, and Bay Trail. Because there’s no way that the A9 is faster than 25W Haswell or even 15W Broadwell CPUs.

          • hasseb64
          • 4 years ago

          Not in transistors, but in productivity probably.
          A Haswell low budget laptop with a HDD is a pain in the ass.

    • Generic
    • 4 years ago

    Being selfish as I am, rather than speculating on their numbers:

    At some point, I’d like to see a review of common tasks and benchmarks running Windows (ARM) on an iPad Pro to allow a more direct comparison to various X86 hardware.

    I understand that it still is not really apples to apples (no pun intended), but a “pro” device is only interesting to me personally if its running Windows.

    Come out swinging Apple. Give us Boot Camp for the iPad Pro.

      • blastdoor
      • 4 years ago

      Windows on the iPad pro would be a disaster. There’s not enough RAM or storage to deal with windows, and there’s no way window would be as well optimized for the CPU as iOS is.

      • chuckula
      • 4 years ago

      [quote<]Give us Boot Camp for the iPad Pro.[/quote<] Not sure how that's going to help since there isn't any version of Windows that's really designed to run on the A9X. Oh, I'm aware that windows-on-ARM exists, but one thing you need to remember about the ARM world is that slapping an OS on an ARM device is nowhere near as simple as slapping a generic Windows/Linux/etc. installation on PC hardware.

      • adisor19
      • 4 years ago

      FYI, M$ killed WinRT to the delight of many. WinRT was initially justified in order to twist Intel’s arm to put out low priced CPUs, which it achieved so it has now been shelved.

      It is clear that the ARM race is now a race to the bottom for all the generic SoC makers with maybe the exception of Qualcom. Only Apple has the resources and the $$$ to make high performance ARM chips that rival Intel’s and given the yearly improvements they made and the insane amount of cash, there is no doubt in my mind that they will catch up with Intel and then some.

      This should scare every single company out there to hear that Apple is dabbling in their wares. Take Tesla and the big auto makers. Mark my words, if Apple is serious about their secret self driving car, they’re in big trouble. That mountain of cash can make miracles happen and what we’re witnessing today with the A9(X) cathing up with the mighty Intel is just the tip of the iceberg.

      Good times ahead everyone !

      Adi

        • sweatshopking
        • 4 years ago

        Mkay. thanks for writing M$

          • adisor19
          • 4 years ago

          Where have you been ? It’s back in vogue, brah.

          Adi

          • September
          • 4 years ago

          M$ should have been their ticker symbol.

        • NeelyCam
        • 4 years ago

        [quote<]Only Apple has the resources and the $$$ to make high performance ARM chips that rival Intel's[/quote<] Samsung has enough $$$ to do that too, not to mention its own finfet process that it can tweak to optimize for their own ARM chips is they so choose.

          • Beelzebubba9
          • 4 years ago

          They don’t seem to have any core design teams in house since I don’t think they’ve ever used a non-ARM core?

          High performance, low latency single thread performance is the hard part and what separates the, err, Intels from AMDs. 🙂

            • the
            • 4 years ago

            [url=http://www.androidheadlines.com/2015/03/samsung-looking-ahead-to-2016-with-their-own-custom-cpu-core-in-exynos-processors.html<]Samsung does have a CPU core design team now so goes the rumor.[/url<]

        • torquer
        • 4 years ago

        Using M$ like its 1999 is really lame.

    • HERETIC
    • 4 years ago

    “Then there’s the bit where Apple says the A9X is faster than 80-90% of mobile PCs shipped in the past year.”
    Reminds me of the case in the UK where apple was brought up for false advertising.
    Their defense was-“But it’s advertising, no one believes advertising.”
    RDF to the max.

      • Atradeimos
      • 4 years ago

      Yeah… But every time they’ve said “2x CPU/GPU performance” in the past, benchmarks show they’re more or less right.

    • Ninjitsu
    • 4 years ago

    I too am curious about what they call a “portable PC”.

    • odizzido
    • 4 years ago

    faster than 80% of mobile PCs?….I guess the first question is how far did their marketing department have to stretch on what people consider a PC in order to get that?…..Old phones maybe?

    Unless of course their like 5W CPU is now crushing over intel’s much higher wattage mobile processors….yeah, I am doubting it.

    • the
    • 4 years ago

    My prediction: SMT. They’re not increasing the core count but the active number of concurrently running threads. For an already wide architecture like Cyclone, SMT is a natural fit considering the probability of used resources increases the wider the design.

    Now there is a case against SMT: power consumption. The more resources that are active, the more energy it’ll consumer and the hotter it will run. However, I’d argue that SMT consumes less power than activating another core of the same design. The resource sharing with SMT is where the power savings will stem from.

      • NoOne ButMe
      • 4 years ago

      Unlikely. SMT increases heat quite a bit for every implimentation I’ve seen. In mobile that’s not exactly a desired goal.

        • the
        • 4 years ago

        But is it less than waking up a second core to do work?

      • Damage
      • 4 years ago

      Interesting thought.

      • blastdoor
      • 4 years ago

      Are there really enough threads running on iOS devices to make SMT worthwhile? Up until very recently Apple has only been making dual-core SOCs and has put a huge emphasis on single-thread performance. The A8X is their first SOC to go beyond dual-core, and even then they just added one more core.

      Even Intel only uses SMT in chips aimed at the high-end.

      If the iPad Pro is very successful, or if Apple starts putting A-chips in Macs, then it will make sense to add SMT. But until then I think it’s much more likely that Apple will get performance boosts from increasing IPC. Intel’s Core chips represent the world’s best estimate of what the maximum feasible IPC is, and Apple hasn’t caught up with Intel yet in that regard (at least not as of the A8). So until Apple catches Core on IPC, I doubt we’ll see Apple go SMT.

        • the
        • 4 years ago

        You are absolutely right that developers have gotten used to the idea of a dual core Soc. Here is where SMT comes in handy with legacy software: run both of those threads on the same core and keep the other core power gated. It is rather rare for two threads to fully load Apple’s SoC so this configuration is more about increasing power efficiency than raw performance. The second core is still there for situations where the extra processing speed is necessary.

        Intel’s usage of SMT is mainly for product segmentation but not necessarily at the high end:
        Pentium/Celeron: dual core
        Core i3: dual core + [b<]SMT[/b<] Core i5: quad core Core i7: quad core + SMT Only the ultra lowend Pentiums and Celerons get by with the ability to run two threads siultaneously. The rest of the consumer line up supports 4 threads concurrently or more. I also disagree with the idea that Intel's chips are the best estimate for what is possible with IPC. IBM's POWER8 beats Haswell in terms of IPC today. There was also the legendary Alpha architecture which was doubling Intel in IPC at the time as well as offering higher clock speeds. The reality is that what Intel offers is some of the best performance per watt which is far harder to do than just raw performance per clock. Intel is incredibly strong here not just because of the x86 design but also due to Intel having the best fabs out there to make them.

    • brucethemoose
    • 4 years ago

    Does anyone remember Wide I/O 2 and Samsung’s plans for Wide I/O 3?

    Apple has disgustingly large margins and a fetish for tiny memory pools, so they’re the perfect customer for exotic, relatively expensive memory like Wide I/O. I know “double the bandwidth” suggests DDR4, but in a world of bandwidth-starved mobile GPUs, a A9/A9X with HBM-like memory would be pretty cool.

      • the
      • 4 years ago

      I think volume production was planned for next year. At least for the volumes Apple tends to order.

    • mad_one
    • 4 years ago

    The claim is interesting indeed. First of all, what’s their definition of mobile PCs? For me that would be anything running Windows, including tablets. I would not count Chromebooks or Android tablets, but who knows what they are doing.

    Given the performance figures, I would expect 3 and 4 cores. If “mobile PCs” are notebooks, they would need to be faster than a dual core 15W i5 (pretty common, even in cheap notebooks). That seems hard to believe, even for threaded workloads where the 2 extra cores should do better than hyper threading.

    If chromebooks and Intel based tablets are factored in, maybe 80% are made of Core-M, Core i3 and various Atom devices. I think they can beat all of those in threaded benchmarks. The mobile i3’s are pretty slow in their 15W variety. Beating Core-M with 1-2 threads would be impressive, but seems unlikely to me.

      • UberGerbil
      • 4 years ago

      [quote<] I would not count Chromebooks[/quote<]I wouldn't either, but in Apple's world I'm sure they consider those "mobile PCs."

        • derFunkenstein
        • 4 years ago

        They’re also not 80% or 90% of the mobile PC market, either. Would be quite a trick, though, to see them claim to have caught Haswell (as I’m sure more than 20% of the market for the past 12 months would be).

        That would put the Core M MacBook at something of a disadvantage against the iPad Pro, which to me seems unlikely.

    • tipoo
    • 4 years ago

    One could point to Atom/other low power PCs like Chromebooks and say that’s what Apple is talking about, but I’m willing to give them more credit than that. The A8X in the Air 2 was already shockingly close to Core M, so if the A9X is near doubling the CPU and GPU performance, it could well be better than Core M laptops as well.

    Not expecting it to beat 15W ultrabooks of course, but beating 5W Core M is within reason.

      • Andrew Lauritzen
      • 4 years ago

      > Not expecting it to beat 15W ultrabooks of course, but beating 5W Core M is within reason.

      Absolutely, particularly if it’s 3-4 cores w/ multithreaded workloads I think this is very doable given benchmarks and their claims. Will be interesting to see 🙂

        • Beelzebubba9
        • 4 years ago

        Yeah I’m super curious if the A9X could come close to Core M, especially in power over time benchmarks. An 80% speedup generation to generation in this day is pretty incredible for a CPU of this complexity.

          • tipoo
          • 4 years ago

          Especially considering the comparison was already to their tri-core, so unless they went with more than 4 this time (unlikely imo), that 80% would be hard to hit. Can’t wait for someone to get their hands on it and see what’s in there, Anandtech has done pretty great deep dives into Apple architectures before.

            • Beelzebubba9
            • 4 years ago

            Yeah it looks like Geekbench scores are pretty even between the Retina MacBook and the iPar Air 2, for whatever that’s worth. This could be very interesting.

            • Andrew Lauritzen
            • 4 years ago

            We really need more than Geekbench and web benchmarks though to make any sort of architectural comparison though. Hopefully someone comes up with something better this time around.

            • tipoo
            • 4 years ago

            There’s always the completely objective, non-cheatable Sunspider 😛

    • ltcommander.data
    • 4 years ago

    Counting Chromebooks as mobiles PCs, wouldn’t they represent a significant portion of mobile PC sales, and using Atom, Celeron or ARM processors bring down the overall performance level of an average mobile PC? That could contribute to Apple’s better than 80-90% of mobile PC performance claims.

    At WWDC this year Apple introduced Bitcode allowing apps to be stored as LLVM intermediates on Apple’s servers so that they can always be compiled with the latest compiler on distribution to users. They said one of the motivations for this is to allow apps to automatically take advantage of new processor capabilities they might be adding. They’ve also went ahead and enabled Bitcode by default for iOS 9 app development. This could enable them to customize the ISA to their requirements like they did with armv7s for Swift A6/A6X and give them more flexibility in how to increase performance than sticking to a strictly armv8 design would.

    • Hattig
    • 4 years ago

    Yeah, it’s either a mix of higher IPC and higher clocks, or there’s an extra core in there and higher clocks. Or an extra core and higher IPC and not greatly changed clocks.

    The GPU must have more shader clusters however to show those gains. I’m guessing 50% more shaders running about 20% faster. However maybe 14nm is super amazing and they can run the same 8-core GPU at nearly 900MHz in the same power budget as the 450MHz 20nm GPU in the A8X.

    The one thing we can say for certain, given the doubled RAM bandwidth, is that it uses LPDDR4 at a decent speed.

      • tipoo
      • 4 years ago

      A8X doubled A8 RAM capacity as well, I wonder what A9X has. If it doubled RAM bandwidth, dare we guess it has more than 2GB?

        • willmore
        • 4 years ago

        Did you mean to say “A8X doubled the A8 RAM capacity”?

        • adisor19
        • 4 years ago

        I’m thinking that’s the case as well. IPad Pro with 4GB.

        Adi

    • blastdoor
    • 4 years ago

    Here’s my uninformed guess — the 70% increase is explained by a combo of 30% better IPC and a 30% clock boost.

      • James296
      • 4 years ago

      with a 10% increase in apple tax

    • windwalker
    • 4 years ago

    Not bad for a “re-branded Exynos”.

    • chuckula
    • 4 years ago

    [quote<]Here's your Apple A9 and A9X confusion thread[/quote<] Thanks Apple!

      • terminalrecluse
      • 4 years ago

      Did anyone else pick up on their quip where they compared the A9X in the iPad Pro to intel’s chips in performance? That leads me to think that they will eventually move their computers to all in house chips.

        • delsydsoftware
        • 4 years ago

        I dunno, Apple played that game when they were using PowerPC based CPUs. The PPC was a great chip on its own (especially compared to a similarly clocked P4), but when the Core2Duo was released, it signaled the death knell for the PPC.

        [url<]http://www.primatelabs.com/blog/2008/06/mac-performance-june-2008/[/url<] They've already migrated from 68k ->PPC ->X86. I don't know if they really want to add ARM to that list.

          • terminalrecluse
          • 4 years ago

          The difference this time being that they are much closer in being able to design it for their own uses having bought people in from PA Semi etc.

          I am not saying that I am for it or against it just that it’s interesting. Secondly, it would add to their margins so I think they’d consider it (not having to pay Intel even given the discounts they must get has got to be compelling). But it would require a collosal rework of software to move to an environment where there wasn’t much in the way of software (who other than Apple and IBM were making software for the PPC at the time?) to x86 which had a lot to choose from and many more developers familiar with the platform. If they did change they’d be doing the reverse and that’s not likely to make the most sense.

          • blastdoor
          • 4 years ago

          Apple is in a better position to fund development now.

            • adisor19
            • 4 years ago

            Yeah that truckload of cash is being put to good use and it shows..

            Adi

          • guardianl
          • 4 years ago

          Apple won’t hesitate for a second to switch if they think they can get away with it. They already have the ARM software ecosystem down with their mobile lineups, and the Macs are small potatoes compared to the mobile profits.

          They’re just waiting until their own CPU designs are good enough, and each year they make pretty good progress towards Intel. 🙂

            • adisor19
            • 4 years ago

            100% agreed. While it will suck to loose x86 compatibility, the masses won’t know the difference.

            Adi

          • derFunkenstein
          • 4 years ago

          Actually, the death knell for PowerPC chips in Macs was sounded about a year before the release of the Core 2 Duo, when Apple announced it was going Intel. The “performance per watt” discussion at the time really confused GEE-off.

          [url<]https://techreport.com/news/8412/apple-confirms-switch-to-intel-processors[/url<]

            • delsydsoftware
            • 4 years ago

            That’s true for Apple, but IBM continued to use and develop the PPC architecture in their data centers for a while. The new generation of Intel’s CPUs basically killed off what was left of the PPC.

          • Kurlon
          • 4 years ago

          I don’t think Apple will move the OS X x86 platform to ARM. They’ll just keep expanding the ARM based IOS platform and shrinking the OS X side until the only options left are IOS on ARM.

        • HERETIC
        • 4 years ago

        Don’t really take a lot to beat Intel race to sleep core M.

Pin It on Pinterest

Share This