It's worth pointing out how extremely far ahead Apple seems to be in terms of CPU power and efficiency. It tends to go under the radar a bit because it's not easy to run your own software on an iPhone, and most apps on these things are not serious workhorse loads (a few specific use cases are, most are not).<p>The Apple A13 - even its implementation in the iPhone SE, in microbenchmarks achieves on par single core performance [1] with the Core i7 8086k [2] and Ryzen 9 3950X [3]. That's the highest single core performance you can buy in PCs <i>in principle</i>.<p>I don't have to explain how insane it is a 5-ish watts smartphone CPU delivers that kind of performance, even if it is in bursts. By sticking to Intel or even x86 in general, there is ample evidence Apple is leaving a lot of performance on the table. Not just in MacBooks - but for the <i>Mac Pro</i> too.<p>[1]: <a href="https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q=iphone+11" rel="nofollow">https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q...</a> - the iPhone 11 with the A13 has to do as a surrogate while the SE is just out, they benchmark the same.<p>[2]: <a href="https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q=8086k" rel="nofollow">https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q...</a><p>[3]: <a href="https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q=3950X" rel="nofollow">https://browser.geekbench.com/v5/cpu/search?utf8=%E2%9C%93&q...</a><p>It's worth noting that Geekbench is a pure microbenchmark. The iPhone will not sustain performance as long as the others. The point is that Apple could solve this when moving to bigger devices.
I remember watching this video on the Intel 8086 by Harvard Business School, and they mentioned how revolutionary it is from a business perspective because it took a vertically integrated market and made it a horizontally integrated market.<p>(I can't find the video after a cursory search...)<p>Is this is the reversal of this revolution? Are we going back to a vertically integrated market due to consolidation in market players or because of performance / power concerns? Everyone seems to be making their own chips and boards these days. Google / TPU, AWS / Graviton, Microsoft / SQ1...<p>Will we ever see a fragmentation in ISAs a la EEE? IMHO that would be a catastrophic regression in the software space, easily a black swan event, if you say needed to compile software differently between major cloud vendors just to deploy.
I know it won't be exactly the same thing as the 68k and PPC eras, but I'm excited nonetheless. I'm a sucker for stories of going against the mainstream and pulling it off. I remember being wholly underwhelmed when Apple switched to Intel in the mid 2000s; my PPC Mac mini was more performant than the first Intel mini by far, especially the GPU. Given how powerful the A series mobile processors are (the new iPhone SE makes my Galaxy Note 10+ look like a slouch in benchmarks), I have a feeling the new Macs will be worth a look for anyone without hard requirements for Windows 10.
This might be OK for consumers and it might be OK for the small business home/lone-wolf developers targeting B2C and small business but it seems like it will be the death knell of Macs/Macbooks being used for a lot of big commerical & enterprise software.<p>A lot of the Mac popularity got a boost by developers loving Macs in the early/mid-2000s because you could target unix so easily compared to a Windows machine and you didn't have to deal with the pain of playing systems integrator to run linux.<p>Today it's a lot worse cause most of us are targeting container technologies on linux. So even if we can compile on the Mac fine we're running a VM (Hyperkit) and taking a performance & memory hit from that compared to a native linux system.<p>If they make the whole lineup ARM and we're now going to be stuck doing an X86 VM on Arm + containers it's going to be even more painful. Given the premium pricing this would probably force a lot of companies to get serious about native linux machines. Painful at first but it'd save a lot of money in the long run. Where I work we already have tried this a little cause the performance on the Mac was starting to suck, you can get stuff like the System76 laptops and get more hardware for less money, there was still just quite a bit of integration pain to use those laptops last time we tried.<p>Yah you can deploy everything from Mac -> cloud when you actually want to run your software but that does make the develop -> build -> deploy -> test cycle take even longer. That aspect is no different whether the mac is x86 or ARM.<p>I'd argue if Apple had stayed PPC all this time and was just transitioning to ARM practically no enterprise shops or startups targeting linux in the cloud would ever be using MBPs, everyone would have found a different solution over the last 20 years. Macs were always just cool curiosities for a lot of software dev until they got to x86.
At this point, it's not interesting unless it's officially announced.<p>We've seen endless wishful thinking and stories on this subject come and go over the years.
The big question here is - are users of these Macs still going to be allowed to decide what software they can run or will that decision be made by Apple like on iOS?
I don't know if buy that Apple would do this for battery life. In certain use cases, we already get 8-10 hours out of their laptops. Of course, running certain apps or heavy loads will shorten that to 2-3 hours, but I can do the same thing on my iPhone, certain apps will chew the battery up in no time. Won't the same thing happen with an ARM MacBook Pro?
What interests me is how many custom chips they’ll pack in to their processor nice they have total control. Phones continue to have more and more specialized accelerators, and I could see computers following that path.
The big question is will they run macOS or will they essentially be iPad-pro books, my guess is the latter, it's not what I'd want but Apple could gain a lot more in terms of control and revenue that way.
It's funny how Microsoft didn't succeed with the push for ARM.<p>Their latest try with 8cx and SQ1 wasn't a big hit. The CPUs aren't the best, MS doesn't want to pour as much money as Apple into CPU research.<p>They also try a gradual move, targeting both x86 and ARM as they don't want to/don't afford to upset consumers.<p>If Microsoft would make next version of Windows ARM only, most software makers would probably target last version of Windows, most users would sit at the last version of Windows and Intel and AMD would happily sell their x86 CPUs.
Hopefully Apple will go back to being a integrated hardware and software company.<p>It makes absolute sense to me for a future MacBook Pro to have a great AMD/Intel CPU and one of their high performance A13X chips with neural accelerators to accelerating specifics tasks to improve performance and battery life.<p>I'm thinking audio/video encoding in FCPX, neural AI assisted tracking, color matching, face detection, ProRes acceleration, H264/H265 acceleration.<p>Would make the price tag of a new MacBook pro much more palatable for FCPX/Logic X users anyway.
I'm not holding my breath, but I really hope Apple works with Microsoft to ensure that Boot Camp still continues to work. Windows on ARM on a MacBook with a powerful A14 sounds like a delight.
I'm not a Mac user anymore, but this might be interesting to move the ARM market on laptops. I'd like to have good options for high performance ARM linux supported laptops.
This would close the gap beetween Ipads and macs even more... after this transiction is done, do you guys think that apple is going to unify both products line?
A related question I was recently asking myself: would it make sense to put two (or more) A13 SoC in a Macbook (context: Apple must see enormous economies of scale now that the A13 is also in the iPhone SE)?<p>Ask HN: Would it make sense to put two A13 into a MacBook or iPad?<p><a href="https://news.ycombinator.com/item?id=22955301" rel="nofollow">https://news.ycombinator.com/item?id=22955301</a>
I wonder if this will have an impact on other parts of the computer besides CPU/GPU - for instance, connecting an iPhone/iPad to Airpods is buttery smooth, but my experience connecting to them on a last year Macbook Air is as crappy as any other Bluetooth device. Maybe whatever they're doing in the mobile line will come to Macs too?
Many suspect that ARM Macs are coming, but I would not trust reporting from “The Big Hack” Bloomberg [1].<p>From the article, Bloomberg reported [2] that Apple would also do it in 2020, so they’re planning on being eventually right.<p>[1]: <a href="https://daringfireball.net/2018/10/bloomberg_the_big_hack" rel="nofollow">https://daringfireball.net/2018/10/bloomberg_the_big_hack</a><p>[2]: <a href="https://www.bloomberg.com/news/articles/2018-04-02/apple-is-said-to-plan-move-from-intel-to-own-mac-chips-from-2020" rel="nofollow">https://www.bloomberg.com/news/articles/2018-04-02/apple-is-...</a>
If MacOS "hides" certain files, prevents mounting `ext4`, requires signing of apps with Apple-ssl-keys to run any binary, tracks every user's app installs and useage, limits installs of applications from only the Apple Store, or any number of other dark-pattern consumer behavior, this loyal Mac user says NO
> Apple Inc. is planning to start selling Mac computers with its own main processors by next year [...] according to people familiar with the matter.<p>Sounds like it’s just tech experts speculating about this. Which they’ve done for a decade or so.
Considering how they came up with Swift, I'm not surprised. They just "get it" with coming up with efficient, yet practical approaches to technical problems.
I’m wondering how Apple got to this position in first place. How does a company that doesn’t specialize in CPU design, beat a company who specializes in CPU design; Intel??
I'm curious when Apple had anticipated this switch. Why deprecate 32-bit in Catalina? Do their 64-bit APIs help them emulate x86 apps more easily?
ARM-based Macs are going to be a pain if they're as locked down as I suspect they might be. But damn are they likely to be fast if the rumours in that article are true. 8 high-performance cores on a next-get chip when Apple's A-series chips are already beating intel chips in single-core performance.<p>I suspect ARM compatibility will be a problem for ~6 until major software houses get themselves sorted out. Most open source software already has ARM support anyway. And of course it's possible that Apple will actually include some level of hardware support for x86 emulation.
this makes me sad because I've just recently got a MacBook Pro, and it was not cheap;<p>still had hopes that it could last >10yrs as my last one did
This is going to be very interesting if true.<p>I'd buy two such ARM mbp on its release, one for the regular macos setup and one for ubuntu linux.
It seems many people wonder why Apple's CPUs does so good compared with x86 and other ARM makers in Geekbench.<p>Here are my guesses:<p>1. Intel is still leading all the rest for single core performance, but it does that at very high frequencies and using a very high power envelope due to process node.<p>If Intel could use 7nm, they could have had a much higher performance if they desired. I can speculate here that they wouldn't have gone for the absolute best performance, because they don't have to. Their goal is to make money, not to make the fastest CPUs they can.
When Intel had a clear lead, they improved the CPUs little by little, each generation.<p>2. AMD has the best performance per dollar and watt and also the best multicore performance. That does bring them a lot of sales and they are happy with that. Their goal is to stay a bit ahead of Intel, so they can sell more CPUs. So their strategy is to be have a better value for dollar. Also, even if they are technically ahead in some areas, there's due to Intel being behind TSMC.<p>3 ARM<p>Arm makes CPU designs and it seems it doesn't care too much about having a clear lead in performance. Maybe with their investment in Ampere and the push towards ARM on servers from the likes of Amazon, they will start to care more. But right now, ARM makes a ton of money by licensing slow and power efficient cores to hundreds of licensees.<p>4 Qualcomm<p>It's not clear if Qualcomm has the know how to push performance. Or that they are willing to allocate the resources for that. I can speculate that they would have done that with 8CX and SQ1 they built for Microsoft if they had the know how and made business sense.<p>Qualcomm doesn't see Apple, Intel, AMD as threats. They compete mostly with Mediatek, Samsung, Huawei. From here, they are very safe. Mediatek optimizes for the absolutely cheapest CPUs they can make, Samsung doesn't care much about CPUs and most of their phones use Qualcomm while Huawei doesn't threaten Qualcomm much yet as their best CPUs are on par with Qualcomm's and they don't sell CPUs to other phone makers.<p>So Qualcomm thinks they can do the same things for the foreseeable future: get some cores from ARM, optimize them a bit, pack with their modem and make a nice sell to the Android makers.<p>The disruption of this model where most CPU makers are happy with slower CPUs can come from ARM server space.<p>If the companies pushing ARM for servers are smart, capable and hungry enough, they might start to eat Intel's and AMD's pies. That would push Intel and AMD to make better optimization and the tech might slip in mobile CPUs, too.
"Apple is exploring tools that will ensure apps developed for older Intel-based Macs still work on the new machines."<p>Part of their release timing might be to ensure that emulating Intel on Arm has acceptable performance for most users.
I always wondered how much of my development environment will keep working on ARM? Electron apps such as VSCode will work fine. Node.js, Python and Ruby will also work okay. Docker can work too. What else?