back to article Apple Arm Macs ship, don't expect all open-source apps to work without emulation – here's what you need to know

Mac hardware based on Apple's M1 chip has started showing up on early adopters' doorsteps, and the machines appear to perform well, even when the 64-bit Arm-based devices are emulating x86_64 instructions using Apple's Rosetta 2 emulation layer. Geekbench scores show M1 Macs outpacing prior Intel-based models by a good margin …

  1. Anonymous Coward
    Anonymous Coward

    Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

    By all accounts Rosetta 2 is doing a great job, don’t forget the M1 is the low end 1st attempt at this and is a direct replacement for low end models.

    Higher end intel gear is still available for the MacBook Pro & Mac mini.

    Power usage in the Mac mini is interesting, 4.2w idle, ~30w peak.

    I’m still awaiting detail on how the 8GB unit performs and if that ssd is actually quick enough to mask lack of ram.

    1. A random security guy Bronze badge

      Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

      I guess that the M1 based systems are not yet ready for full blown software development outside of what Apple provides. Having no Virtual Machine support makes most web app testing pretty hard; sometimes you really want to run software in your own boxes instead of deploying to the cloud and then trying to debug/test it.

      I am actually hoping that Virtual Box is ported. And they provide a more powerful system for the 16” system.

      Maybe a 16 core, 32 gb system. Who knows.

      1. GiantKiwi

        Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

        Eh?

        VM's on Linux boxes are still a thing, you know.

        1. Gonzo wizard
          FAIL

          Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

          I should get a second machine (that isn't ARM) so that I can (a) run Linux and then (b) deploy a VM there? Instead of, maybe, just firing up docker on the machine I bought and paid for?

          There are so many reasons why what you're suggesting won't fly. My last three clients wouldn't let me (freelance) or their own staff do exactly this, yet they all equip their developers with Macs.

          In any event getting a VM on company infrastructure is in my experience very time consuming and must inevitably be justified. Additional physical machines aren't just lying around, need to be configured for the local network, etc etc etc... it's a high friction situation. I just wouldn't bother looking at an ARM based Mac until I can do everything I need on the machine in front of me (which I can now on my 2015 Pro).

          Your comments are ill-considered, I'm sorry to say.

          1. Gonzo wizard

            Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

            Any thumbs downers care to respond? I’m genuinely interested in debating this and understanding the other side of the argument.

      2. Anonymous Coward
        Anonymous Coward

        Having no Virtual Machine support

        Yet. Both VMware and Parallels are expected to release versions soon.

        1. Anonymous Coward
          Anonymous Coward

          Re: Having no Virtual Machine support

          Care to explain the thumbs down?

          1. mevets Bronze badge

            Re: Having no Virtual Machine support

            I think it is a way of expressing disagreement, but could be general dislike.

    2. Kevin McMurtrie Silver badge

      Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

      Don't forget that Apple reserves something like 20% of RAM for compressed memory. This works well when apps have large amounts of idle memory, but it causes performance draining page faults for data crunching apps where all the memory is active. Figure there's less than 10 GB of full speed RAM left after the OS, hardware, and swap cache take their cut from a 16 GB system. An 8 GB model might have under 4GB of the good stuff.

      1. Dave 126 Silver badge

        Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

        >Don't forget that Apple reserves something like 20% of RAM for compressed memory.

        That's how macOS worked on Intel - does it work that way on these M1 chips?

        1. Gonzo wizard

          Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

          The system RAM isn't shared with the graphics hardware on existing MacBook Pro models, I can't speak for the compressed memory thing.

    3. Dave 126 Silver badge

      Re: Still a lot of work to be done porting, but Rosetta 2 is picking up the slack

      > Power usage in the Mac mini is interesting, 4.2w idle, ~30w peak.

      Just to be clear, as Anandtech tested it, that's the power consumption of the entire Mac Mini (including its internal power transformer) since they weren't allowed to take the unit apart. They estimated that the peak of the M1 CPU GPU SoC is under 25W, after the power supply losses are taken into account.b

  2. 45RPM Silver badge

    My initial thought was a bit of a grumble. It won’t run Linux, I thought (I can accept that it won’t run Windows - and I don’t really care. Windows isn’t part of my general use-case, and I’m prepared to bet that it isn’t part of the use-case for 99.5% of Mac users). But then, I thought, most users won’t care if it can’t run Linux either - and, to be honest, the only OSs installed on any of my Macs are OSs from Cupertino. I have other computers dedicated to Linux or Windows.

    This isn’t exactly a new situation, or thinking different either. This is a return to the way things were - albeit with one major improvement. In the 70s, 80s and even to a certain extent the 90s, a significant percentage of the computers that you could buy were like this. If it had an Acorn, Apple, Commodore or Atari badge (i.e. the big players that we all remember and love) then it would run the OS that its manufacturer wrote for it, and no other. Well, no other unless you were prepared for significant jiggery-pokery.

    The major improvement, of course, is that in those days the CPU was still an off-the-shelf part, and the OS and CPU were not designed one for the other. Silicon is just as much designed for macOS as macOS is for Silicon - that phenomenal speed is not magic. It’s synergy. Perhaps that synergy, more than anything else, is the significant breakthrough here.

    I don’t think that there’s any need to worry about software support. Even if most of the software doesn’t get recompiled, and remains running in Rosetta, these new Macs are still up there with the best that Intel has to offer for general computing purposes. But that’s not going to happen - there’ll be plenty of support, and quickly. Apple has made the transition about as easy as it could possibly be. And, of course, if you don’t need mind-boggling speed but you do care about being able to run whatever OS you choose then this may not be the computer for you - but, luckily, AMD are making some tasty and quick CPUs that will be right up your street.

    Would I buy one though? No. This Intel-slayingly quick processor is Apple’s ‘budget’ model. This is the low end - and a first attempt at that (historically, of course, first stabs are always compromised - Apple's, perhaps, more than anyones). I want to see what they can do with the high end. When the high end models launch, I’ll open my wallet.

    1. Dan 55 Silver badge

      If it had an Acorn, Apple, Commodore or Atari badge (i.e. the big players that we all remember and love) then it would run the OS that its manufacturer wrote for it, and no other. Well, no other unless you were prepared for significant jiggery-pokery.

      IIRC all of them let you run e.g. some Unix variant and some had PC compatibility. The 8-bits before them let you run CP/M if they had a Z80 CPU. None of them made you jump through as many hoops as Apple does today.

      1. Dave 126 Silver badge

        Things like the Amiga had various custom chips as well as the 68000. The OS was compiled for the specific version.

        You would have to go through a lot of jiggery pokery to get another OS running on it as smoothly as the one it shipped with.

        1. Dan 55 Silver badge

          Amiga UNIX (Amix) was a thing, although Commodore priced it too high. Sun wanted to resell the A3000UX and Commodore screwed up there too.

          Linux and NetBSD have Amiga ports, NetBSD for the Amiga came out in the mid 90s.

          Many UNIX programs could be ported to AmigaOS as just one more #ifdef in the list of #ifdefs.

        2. 45RPM Silver badge

          Whoops. That’s a really good point. I should have thought of that.

      2. Rainer

        Ran NetBSD arm32 on my RISC PC 600.

        Was fun downloading 20-ish floppy disks at uni (and re-downloading a couple of them because the drives in the old Apollo workstations were flaky as hell).

        Didn't really know what to do with it, though, at the time.

      3. Gonzo wizard
        Thumb Up

        This...

        And this is the thing. Today's software may well appear, but having slowly but surely constructed a walled garden there will inevitably be software that Apple won't want on their platform. I'm a current Mac owner but I'm still on Mojave, and won't be buying any more. I'll give my money to a generic manufacturer with a supported Linux build.

        I'm happy to concede that many people will be perfectly happy with the new platform. I'm just not one of them.

        1. 45RPM Silver badge

          Re: This...

          Thumbs up for a well made point.

          I hope that you’re wrong - because this is the huge difference that separates macOS from iOS. That said, I do like good fit and finish. So I’m going to stick with Apple for a while longer - and I’ll desert the ship when the garden wall gets so big that I can’t jump over it any more.

      4. Tim99 Silver badge
        Coat

        Just to troll you

        "IIRC all of them let you run e.g. some Unix variant..."

        macOS is a Certified UNIX, LInux ain't: opengroup.org. OK it's more complicated than that.

    2. Jan 0 Silver badge

      @45rpm

      > In the 70s, 80s and even to a certain extent the 90s

      > Acorn, Apple, Commodore or Atari badge (i.e. the big players that we all remember and love)

      But, apart from a smidgen from Commodore and a tad from Apple, those were toys. "Big players": business users and programmers were using DG, DEC, National Semiconductor, Sun and so on. Look how quickly they all started with, or took up, Unix if they survived into the 90s.

      Not that this aside really matters one jot to Apple as long as Apple can retain and grow customers.

    3. Soruk

      Linux on ARM started life as a port to run on Acorn kit. I remember seeing it run on an Archimedes machine. Okay, it was slow as hell, but seeing it able to do stuff on an 8MHz box was an achievement.

      (And knowing how well RISC OS ran on it, it still feels blisteringly fast on the original Raspberry Pi.)

    4. Anonymous Coward
      Anonymous Coward

      I suspect you will fins that there are quite a lot of Mac owners, like me, who would not have bought one without the ability to run Windows.

      Not the mac zealots granted, but for me - at least 5 years ago when I bought it - the hardware of a Macbook was a lot nicer than the OS.

  3. Dave 126 Silver badge

    Clear communication

    Applications running under Rosetta 2 run faster than they do natively on the Macs these new M1 models replace. That likely isn't a coincidence, but the result of a deliberate decision at Apple to only release the M1 Macs at this time. It makes their communication, their marketing if you will, that bit easier for them. Competency.

    Hats off to whoever planned that, as well as hats off to the silicon design teams (who apparently worked closely with the Rosetta 2 team to make this possible.)

  4. Dan 55 Silver badge

    Geekbench

    I hope it's improved since 2013 when Linus said "Geek Bench has apparently replaced dhrystone as your favourite useless benchmark" and other less complimentary stuff. Linus Torvalds, that is.

    1. Dave 126 Silver badge

      Re: Geekbench

      There's a lot of distrust of Geekbench, as you note. That's why Andrei over at Anandtech appears to have been throwing everything he can think of at it since taking delivery of an M1 Mac Mini. There should be enough there to give people an idea of how well it will perform for their real life work loads.

      Quite a few typos in his report suggest that he's either not slept yet or has just wanted to post the report ASAP:

      https://www.anandtech.com/show/16252/mac-mini-apple-m1-tested/2

      1. TheRealRoland

        Re: Geekbench

        Having read a bunch of those comments, i like the fight between three camps - roughly saying "this is not an apples to apples comparison, you should wait with benchmarking until the other manufs have caught up!" and "the benchmark software is suspect" and "the other machines are not on the same wattage, so how can you compare?!"

        That thing seems fast.

        But, not in the market for it. I just got my pi4, all good for now! ;-)

        1. DS999

          Re: Geekbench

          Yeah the comments read like a lot of people trying to make excuses for Apple's phone SoC with a couple extra cores is able to match and sometimes beat Intel and AMD's fastest.

          It will be interesting to see what Apple comes up with in a couple years for the Mac Pro. That will really reduce the places they have to run to and say "yeah but what about multithreaded stuff" and compare with 8 and 16 core x86 chips.

  5. CAPS LOCK

    Well, it's nice of El Rego to provide us with gushing quotes from other web sites, but...

    ...since any criticism of Apple can result in excommunication I feel unable to trust web sites that can get access to Apples stuff.

    1. Dave 126 Silver badge

      Re: Well, it's nice of El Rego to provide us with gushing quotes from other web sites, but...

      Private citizens have been taking delivery of M1 Macs for nearly two days now - not just websites who may or may not have received review units. They're in agreement.

      You're inferring that Anandtech are merely shills. There's nothing to stop you reviewing their methodology and reading up on their history before posting here.

      I agree scepticism is generally good thing, but it doesn't stop one from looking at further evidence. That's the definition of cynicism.

      1. Kristian Walsh

        Re: Well, it's nice of El Rego to provide us with gushing quotes from other web sites, but...

        I do agree, but if you take Anandtech (the site that first diagnosed the pathetic self-grounding antenna "design" on iPhone 4, if I recall) out of that list, the criticism becomes a lot more valid. Verge and Engadget are pretty lightweight when it comes to actual technology and have habit of parroting the press release or falling into the "bigger number = better" trap, while TechCrunch is the "Supermarket Shopper" of the tech journalism world.

    2. Anonymous Coward
      Anonymous Coward

      The real problem is apple getting access to their stuff not them getting access to apples stuff

      These new macs appear to be phenomenal user surveillance pieces of equipment, they might not be fully there yet they do have the potential to do amazing things in the area of user surveillance.

      1. Dave 126 Silver badge

        Re: The real problem is apple getting access to their stuff not them getting access to apples stuff

        What, why? Because they're fitted with better than average microphones? Because they don't crash every half hour like my Windows 98 PC did?

        My sharp paring knife has the potential to do amazing things in the area of causing blood to come out of my fingers. But that's irrelevant because it's very good at slicing tomatoes.

        1. el kabong

          The surveillance is done for the users own good

          Apple has its users best interests in mind, keeping the users protected and away from undesirable influence is top of mind at apple, by keeping them permanently under surveillance apple is able to ensure that they do not stray from the pack and get lost.

      2. Anonymous Coward
        Anonymous Coward

        Re: The real problem is apple getting access to their stuff not them getting access to apples stuff

        @A/C

        Citation Needed

        Cheers… Ishy

        1. Anonymous Coward
          Anonymous Coward

          Re: The real problem is apple getting access to their stuff not them getting access to apples stuff

          Well it seems Apple have now reached parity with Win10 when it comes to intrusive and all seeing monitoring of all your computer use activity.

          https://macintouch.com/homepage/post/9881/apple-remotely-disables-many-macs/

          If you dig around it seems that unlike Win10 its almost impossible to lock down Big Sur without breaking it to stop it phoning home.

          Its all for your own good, you understand... Apfel Macht Frei.

          So just another piece of very creepy trojan software just like Win 10 and Facebook anything.

          1. Anonymous Coward
            Anonymous Coward

            Re: The real problem is apple getting access to their stuff not them getting access to apples stuff

            I guess the proves that we’re only about ever three posts away from a Hitler comparison.

            Read more deeply on the OSCP issue and you’ll learn that the sky isn’t falling in quite the fashion you imply. But why bother with facts... just keep on RTing those tweets from Donald eh?

  6. Anonymous Coward
    Anonymous Coward

    I'm going to be called a Fanboi but...

    ...I feel "the machines appear to perform well"

    is possibly underselling the achievement of double the battery life combined; "outstanding", "blows away every ultraportable, with no fan noise to get in the way." "the future of CPU design" & "over-performer."

    1. theOtherJT

      Re: I'm going to be called a Fanboi but...

      Maybe just being cautious. Some of these benchmark figures appear too good to be true - and when things appear too good to be true, they usually aren't true.

      I really hope that the Apple silicon is everything Apple claim it is, because regardless of how I feel about Apple (and dear gods do I _hate_ Apple) competition is good. If they can force the x86 boys to up their game then so much the better, but until I've had a couple of weeks to collect a good spread of numbers from a bunch of different sources I'm going to remain skeptical.

      Remember how good Netburst looked on paper when it was first debuted before we all found out what the thermal performance was like and how despite the unprecedented (for the time) clock speeds, everything tanked when the branch prediction failed?

      1. Dave 126 Silver badge

        Re: I'm going to be called a Fanboi but...

        Naturally, extraordinary claims merit extraordinary evidence. But given the billions of chips from this family in use by consumers (in iPhones), I don't think there is much room for any major flaws to hide, other than in the Rosetta area. Caution or scepticism at this stage doesn't hurt Apple, they have accounted for it - TSMC can only make so many M1 chips, and these new Macs Apple have released are clearly designed to assuage people's reasonable doubts.

        A month ago an Autodesk employee in their forum posted that whilst there isn't an ARM version of Fusion 360, he noted that it ran perfectly happily on their developer unit under Rosetta 2. Over the next few weeks and months, potential buyers will get a clearer idea of how the software they use behaves under Rosetta 2 from first adopters, and of developers' plans to release an ARM version.

    2. Anonymous Coward
      Anonymous Coward

      Re: I'm going to be called a Fanboi but...

      Rather begs the question why Microsoft didn't go about it this way with the Surface / RT etc. They could have buried x86 by now and caused Apple a headache in the process. Instead we've all had to suffer the best part of another decade of Intel mediocrity.

      1. 45RPM Silver badge

        Re: I'm going to be called a Fanboi but...

        Simple. They don’t have any expertise in CPU design. Apple’s been dabbling in this pool since the days of PowerPC, and it’s been getting really good at it since the iPhone 4 was launched.

        In fact, Microsoft doesn’t have much experience at hardware design full stop (their electronics aren’t in-house, I think, being based on reference designs from Intel and Qualcomm), and Apple has been doing this since day one.

        1. Soruk

          Re: I'm going to be called a Fanboi but...

          I am no fanboy, but Apple have been part-owners of ARM since the Acorn days, thus most likely had more close involvement in it than pretty other manufacturer (maybe aside from Broadcom, who acquired Element 14, which was previously known as.... yup, Acorn.)

          1. Kristian Walsh

            Re: I'm going to be called a Fanboi but...

            Apple sold its shares in ARM in the late 1990s when it needed the cash to keep the lights on. ARM is wholly-owned by nVidia today, having recently been sold by SoftBank, which had been sole owner since 2016 - the same year Apple launched its first (and very much "me too") A1 System-on-Chip.

            Apple does what every other ARM customer does: they licence the CPU cores, and tweak the logic around those cores to produce a system-on-chip best suited to their needs. Then they get another company to deal with the problem of fabricating the resulting chip. In Apple's case, that company is TSMC, and pretty much all of the M1 chips low power wins are down to TSMC's brand-new, industry leading 5 nanometre manufacturing process. Apple won't hold that lead for long, as TSMC does work for everyone.

            However, Apple's engineers can take credit for the good benchmarking results on tasks that require high memory throughput - that's one of their optimisations over a "typical" ARM SoC.

            Until we get other SoCs on the same process, it's hard to say for sure how much of this brilliance is from Cupertino, and how much from Taiwan.

            1. Dave 126 Silver badge

              Re: I'm going to be called a Fanboi but...

              > Apple does what every other ARM customer does: they licence the CPU cores, and tweak the logic around those cores to produce a system-on-chip best suited to their needs.

              It's a bit more than just 'tweaking'. See Anandtech for the details.

              1. Kristian Walsh

                Re: I'm going to be called a Fanboi but...

                Forgive me for not gushing - I worked for Apple for a few years, so I know better than to take anything they put in a press-release at face value, but for the record, I think M1 is a significant leap forward for desktop CPUs and for the ARM ecosystem.

                I read the Anandtech article. Apple have not modified the cores, but they have done a lot of interesting work with the memory interfacing, and it shows in the benchmarks. The relatively ordinary performance on crypto tasks (which are compute limited) shows that the performance gains are due to a huge improvement in memory I/O.

                M1 is the first ARM design that can actually compete with traditional desktop CPUs, and Apple has done some really good work on fitting this device for laptop and desktop workloads. By comparison, Qualcomm's design for Microsoft's Surface ARM product was still in the mindset of a "mobile device" SoC, and that seriously restrained its performance (also, Windows buyers are far more cynical and change-resistant that Apple customers). Apple's approach was to ditch the UMTS support, and use the power extra budget to seriously improve performance, and finally make an ARM a viable alternative to an x86 ISA system.

                However, while I think it's a big deal, I'm not willing to play along with the usual game of ascribing all the brilliance of a new product to the company whose brand ends up on the package, when much of it derives from the work done by suppliers - in this case ARM itself for the CPU cores, and TSMC for the new manufacturing process. Apple is the first customer of TSMCs brand-new 5 nanometre process, and that feature-size reduction has a huge part to play in the M1's impressive performance-per-Watt measurements.

                But there's one final thing that struck me about the reaction to M1 - everyone's talking about how Intel will react, naturally enough, but it's also worth considering what the other ARM SoC vendors might do, too. The closely-related Apple A14 has similarly good benchmark performance to M1, and yet that A14 isn't the runaway leader in the mobile world that M1 appears to be on laptops, and generally the lead swaps between whoever has the newest device on market. If M1 lives up to the benchmarks, what odds on seeing true desktop-focused ARM SoCs from the likes of Samsung and Qualcomm before too long?

    3. doublelayer Silver badge

      Re: I'm going to be called a Fanboi but...

      There are a few things that will only get answered definitely when people have received units and stress-tested them for several days, and those variables might change the experience for some types of users. Such variables include things like whether the fanless nature of the MacBook Air makes the processor throttle performance often and whether doing this leads to degraded performance for certain use cases, whether there are times when emulation is required but it doesn't work right, or whether the shared GPU uses a lot of memory which is more limited on these chips. Some of these things can't really be determined in a numeric benchmark, and existing reviews seem not to have tried a large number of use cases. While these provisos remain, the reports I have seen thus far are promising.

  7. Binraider

    I absolutely loved the original Archimedes. BBC Basic on demand, easy access to disassemblers, a clean, colourful and very responsive UI with a large proportion of it sat in ROM, and a good software selection too. In it's day, the only thing it really missed out on was having the dedicated graphics-mangling silicon that made the Amiga. Acorn got an awful lot right in the architecture of the machine. Acorn got an awful lot wrong by targetting it purely at schools; as they were excellent productivity machines. And low and behold that architecture lives on giving benefits over the competition.

    Not many have commented on another feature of these machines. The price-point. The laptops are probably about right. MacMini slightly over-egged.

    Apple have spotted an opportunity to create decently (if not high) performing lightweight machines using an ecosystem that they've already developed. I won't be picking up the "mark 1" version; these eyes prefer a larger monitor, and, you know what happens to early adopters. Whether the ARM scales to higher performance levels is another matter. More parallel units obviously possible; but what about out-and-out raw performance that the Mac Pro brigade will demand? PCI-Express interfaces for big GPU's etc. I'm hopeful they can pull this off. Server-grade ARM hardware with some aspects of these features have been around a fair while so certainly possible.

    1. Dave 126 Silver badge

      > The price-point. The laptops are probably about right. MacMini slightly over-egged.

      Yeah, that was my initial thought when I saw the Mac Mini price - if Apple sold it cheaper they could shift loads, and after all, there's not much to it beyond the M1, no screen or keyboard. But then I remembered that TSMC's production capacity for 5nm chips isn't unlimited.

      > Whether the ARM scales to higher performance levels is another matter.

      Shouldn't be a problem. Single core performance beats almost everything. Machines built for better thermal performance should allow more cores to run fast, though Anandtech note that a single core can saturate the memory so using more than one core can actually degrade performance.

      The GPU question is interesting because the M1's GPU beats all integrated graphics and many lower end discrete graphics at far lower power consumption. Apple's roadmap doesn't show any AMD GPU's on ARM Macs. This suggests that Apple are confident of building a discrete GPU of their own.

    2. rg287 Silver badge

      Whether the ARM scales to higher performance levels is another matter. More parallel units obviously possible; but what about out-and-out raw performance that the Mac Pro brigade will demand?

      A fair point, and I suspect the Pro will be the last Mac to lose Intel (Xeon) options. The ARM USP is power, so no surprise that it's predominantly focussed on laptops at launch.

      That being said the single-thread performance seems to be up there with AMD & Intel, which implies that a TDP-no-object desktop die packed with just high-performance cores (as opposed to the 4+4 bigLITTLE architecture of the M1) could be quite the thing to behold.

      Of course, Pro buyers really won't tolerate soldered RAM or SSDs and are probably going to want PCIe slots, so the question is whether their blinding performance is just down to CPU design (which is undoubtedly very good) or if it's riding off things like blistering unified memory speeds - can they maintain their performance advantage on a more traditional/modular motherboard with conventional DIMMs and PCIe memory & expansion cards?

    3. theOtherJT

      IMHO it was RiscOS that really made those things. It was light years ahead of anything else available at the time. I mean in 1987 your options were what, TOS, System 5, Windows 2.0, amigaOS 1.4?

      That being said, I have no idea how many of the capabilities of RiscOS were down to the underlying architecture and wouldn't have been possible on other designs of the era.

      1. Binraider

        RiscOS and the ARM architecture were pretty closely linked. Early archimedes examples didn't actually have the system-on-chip design that rapidly became the norm - the cpu space on the motherboard had a little riser and a second board with several chips on it. It didn't take long for the silicon on that board to be integrated, with the attendant benefits of compacting it down.

    4. DS999

      Making a CPU core that runs single thread code fast is an order of magnitude harder that making a bunch of cores work well together for multithreaded code.

      From everything I've read, Apple plans to scale their GPU all the way up to the Mac Pro - there are no plans to support ANY third party GPUs. I'm sure they wouldn't make such a decision if they didn't already know their GPU could scale - and may well have something like that already running in some secret Apple lab deep inside the spaceship.

  8. Dave 126 Silver badge

    Adobe

    Adobe has released a feature-complete Photoshop ARM beta for macOS, whilst stating that they don't officially support Rosetta 2 (although it seems to work). Across various Mac forums this has only reawakened user criticism of Adobe, for their slowness in adapting, their licensing terms and longstanding UI niggles, with much discussion of viable alternatives such as Affinity.

    It's not as if Adobe haven't been slowly porting / rewriting Photoshop to ARM (iPad) for a few years now, or that Apple won't let them have development kits, or that they don't have the resources.

    I'm a Windows user, and as such it reminds me of how slow first Microsoft was to support high resolution monitors properly, and then after they fixed it, how slow (years!) Adobe were to have Photoshop work at high resolutions without the tool palettes being unusably tiny. Grrr.

    Okay, maybe I'm anti Adobe today because for the love of dog I don't understand why in Illustrator it's impossible to have the scroll wheel tied to zoom in / out as it is in Photoshop and in every CAD package from the last two decades.

    1. Gordon 10 Silver badge

      Re: Adobe

      I think that there was a phrase in the article that was the perfect summary.

      "Adobe, historically known for being out of step with <$insert_anything_here> "

      1. Dave 126 Silver badge

        Re: Adobe

        I mean seriously, Adobe can't be lacking in resources given their subscription model. What are they doing with it all? Is there something I'm missing, such as the Photoshop code base being peculiarly Byzantine? Have they just been too complacent for too long, or is that oversimplifing things? Anyone seen any articles featuring interviews with ex Adobe employees?

        1. WolfFan Silver badge

          Re: Adobe

          They don’t care. They don’t have to. They’re Adobe.

        2. Joe Gurman

          Re: Adobe

          Erm, the subscription model means they can throw any old thing they want out there any old time they feel like it. How does it help their bottom line to rush?

          Actually producing usable code during five months of beta versions is for lesser mortals.

    2. Kristian Walsh

      Re: Adobe

      Affinity Designer.

      Seriously. Try it.

  9. deadlockvictim Silver badge

    Rosetta El-Reg-style

    El Reg» El Reg isn't allowed within 100ft of Apple review hardware for reasons you can guess...

    You need to take a leaf out of Apple's book: Rosetta.

    El Reg publishes articles that conform to Apple Marketing (aka propaganda) and you provide a Rosetta-Stone-like icon (e.g. the one from Civilization I) that translates into Regspeak. That way El Reg slowly becomes the most favoured altar boy while we get the gossip from the presbytery.

  10. karlkarl Silver badge

    "Apple Arm M1 Macs ship... tho don't expect all open-source apps to work"

    Why would open-source "apps" not be ready for the new architecture. If we are provided with a C and C++ compiler, it is as simple as just building the source as usual. Since the API hasn't changed, there should be very few platform specific parts needing changing.

    Especially since many of the x86 specific code has already been eliminated whilst "fun porting" to the Raspberry Pi. Open-source is really good here.

    The main reason why iOS and Android are "hard" is because of the terrible toolchains and focus on vendor specific languages. And even then they aren't hard, people just lack interest for such short-lived platforms.

    1. Anonymous Coward
      Anonymous Coward

      @KarlKarl

      iOS and Android are"short lived'?

      Cheers… Ishy

      1. karlkarl Silver badge

        Yes. Android 4.x is pretty much an absolute no go and yet that still came out many years after Windows 7 which still has a large market.

        I don't know how technical you are but if you also look at those platforms in terms of programming APIs, you will see that they generally die and a new one takes there place almost yearly.

    2. theOtherJT

      I'm less concerned about if you can port foss apps to the M1 architecture than if Apple will let you within the OS.

      They really are going out of their way with macos to try and prevent people running apps that don't come from developers with Apple signed developer certificates.

      1. Dave 126 Silver badge

        Really? How do?

        1. Dan 55 Silver badge

          As of Catalina Gatekeeper will not allow you to run an unsigned downloaded binary package or downloaded old-style ELF binary and it's only possible to disable that check with the terminal.

          Signed software is checked online with Apple's OSCP server but this seems completely unnecessary if there's also an XProtect check which is regularly updated and Apple can feed it a rescinded certificate (see recent HP printer driver story).

          Grokking Gatekeeper in Catalina

        2. Dan 55 Silver badge
          Devil

          Just found this:

          Checks on executable code in Catalina and Big Sur: a first draft

          Unlike unsigned Intel code, unsigned ARM code is blocked from running, you must generate an ad-hoc signature first. Wow.

          Sounds like homebrew and other package managers are going to have their work cut out.

          1. karlkarl Silver badge

            "Sounds like homebrew and other package managers are going to have their work cut out."

            They probably just won't bother. There is no technical challenge for them, just old-fashioned bureaucratic / licensing stuff.

            This whole era is going to be a waste of time like Windows RT. Sure, the typical non-technical Apple idiots might just think it is because "Wahh, ARM is hard to program for". When really it is all the artificial crap that Apple imposes. Apple will of course blame ARM rather than their own criminality and the plebs will be none the wiser.

            More shite for the landfill basically.... and ZERO progress.

            1. Dan 55 Silver badge

              And if they do apply for a dev cert, they'd become responsible for everything in their repo. So just one fool would need to complain about homebrew distributing youtube-dl and Apple would revoke everything.

              1. karlkarl Silver badge

                Exactly. It is a bad design and the only correct move is to not play.

                Apple is so... boring.

                1. Snapper

                  Try not to be so stupidly negative, it's so boring!

    3. Def Silver badge

      Well, given that Arm has weaker memory ordering than x86 a simple recompile will be fine until you start communicating between different threads. Then a lot of code that works on x86 could suddenly stop working in strange and mysterious ways.

      Which begs the question how well does Rosetta really handle this?

      1. Dave 126 Silver badge

        > Which begs the question how well does Rosetta really handle this?

        It's an interesting question, but one people will seek answers to after they ask "Does the software I use work on it?"

        Apple have stated that the M1 team and the Rosetta team worked closely together (which has a bearing on your question), and Apple is well motivated to ensure it works well.

        1. Def Silver badge

          I'm a little skeptical that this isn't an impossible problem to solve through recompilation alone though. I don't see how it would be possible to detect missing memory barriers that would be required to ensure thread synchronisation happens correctly on Arm.

          Unless there's a way to disable memory write reordering on a per application basis, of course.

          1. Anonymous Coward
            Anonymous Coward

            A lot will come down to if there are any instances of undefined-behaviour within the code. If there are (and there probably are, as it's tricky not to end up with it accidentally without the use of decent static analysis tools), then it will be down to luck if it works.

            1. Anonymous Coward
              Anonymous Coward

              "thumbs down"?

              Undefined-behaviour means there is a latent defect within the software; it may work "now", but is not guaranteed to work if the code is recompiled with a different compiler (including a different version of the same one) or the settings (generally optimization) are changed.

              Instances of unspecified behaviour, such as the order of evaluation, can also "break" code in a similar way. For example, what is the result of the following?

              i = 1;

              x = i + i++

              1. Kristian Walsh

                Re: "thumbs down"?

                The result of that code is a failed code review :)

                This is a classic sneaky C interview question, because it's actually undefined in the language specification, and so the correct answer is "don't do this". gcc says 3, which how a right-to-left evaluation would work it out, but an argument can be made that the answer should be 2 if you treat it as LOAD x <-- i; ( LOAD A<-- i; INCREMENT i; ) ADD x <-- A;

                GCC throws a warning if you ask it to be picky (you should always build production code with warnings on): operation on i may be undefined.

            2. Richard Pennington 1
              Boffin

              Concurrent processes

              As soon as you start looking at communication between threads, static analysis tools are insufficient (because it ain't static...).

        2. Anonymous Coward
          Anonymous Coward

          Autodesk Fusion 360 which was native on x86 macOS, is reported by Autodesk to run on their Apple ARM developer kits with no noticeable loss in performance.

          If I were a Mac user, that would have my attention. However, most CAD software is Windows only, so therefore so am I.

          I'm guessing here, but maybe the porting of Fusion 360 to OSX a decade ago was taken as an opportunity to clean up the code. Or maybe it was well behaved because a fair bit off it is GPU heavy, and the macOS graphics API Metal hasn't changed in this transition.

      2. Anonymous Coward
        Anonymous Coward

        Really? Doesn't x86 already have pretty weak memory consistency rules?

        1. Def Silver badge

          No, quite the opposite. Memory writes on x86 are never reordered.

          So the following code running in two different threads will Just Work (as far as memory coherency is concerned):

          Thread 1:

          result = somevalue;

          done = true;

          Thread 2:

          if( done )

          {

          // do something with result.

          }

          On Arm, however the processor is free to reorder the writes in thread 1 such that done can be updated before the result has been written, thus breaking the natural order of the universe. To avoid this a data memory barrier instruction must be inserted between the two writes to prevent such reordering.

      3. DS999

        Apple's CPUs support x86's stronger memory model

        Which I'm pretty sure was added specifically for running code that was translated by Rosetta 2.

        1. Def Silver badge

          Re: Apple's CPUs support x86's stronger memory model

          Apple's developer documentation on the matter suggests otherwise:

          https://developer.apple.com/documentation/apple_silicon/addressing_architectural_differences_in_your_macos_code

  11. dave 93

    Universal binaries do contain native M1 code, obvs

    "Microsoft has released a Universal build – containing x86_64 and ARM64 binaries – of its Mac Office 2019 beta. But there's not yet an M1-native general release of Office."

    This is wrong - Universal binaries contain M1 native code AND native Intel code - that's the whole point...

    1. Dave 126 Silver badge

      Re: Universal binaries do contain native M1 code, obvs

      The distinction here is that a Beta is not a General Release. So at the moment there is:

      -a Universal Binary Beta

      - an x86_64 General Release of Office.

      But no M1 Native general release yet, as was stated.

  12. MarkET

    ...because R depends on having an Apple Silicon-ready Fortran 90 compiler

    One step forward, one step...

    1. Francis King
      WTF?

      Re: ...because R depends on having an Apple Silicon-ready Fortran 90 compiler

      Fortran is still a very strong language. Not sure what you're trying to say here.

  13. Jason Hindle Silver badge

    From a developer perspective, it's going to be a very mixed bag.

    It looks like Rosetta pre-compiles intel code rather than translating during execution. I've seen a couple of YouTube vids where the first attempt to lunch an intel app takes quite a bit more time than subsequent launches. I've also seen a demonstration of VS Code running perfectly well. As such, I'd expect intel versions of things like Git and Maven to work perfectly well. That said, for the developer it's going to be a mixed bag:

    - Native versions of Java are here already: https://adtmag.com/articles/2020/11/12/azul-supports-apple-silicon.aspx

    - But no idea when Homebrew will be available.

    For my work, I would want to know that there are either native versions chromedriver and geckodriver or that, at the very least, the intel versions are ok and will play nice with whatever Firefox and Chrome is on the box.

    One interesting point to note: The review units Apple sent out all appear to be base, so 8GB and it doesn't look too shabby at all (and might be good enough for most of the users that don't need to run up virtual machines). Speculation on various fora that Apple upped their memory management game.

    Overall, I'm seriously thinking of picking up a base Air to play with.

    1. Dan 55 Silver badge

      Re: From a developer perspective, it's going to be a very mixed bag.

      From a developer prospective it seems you'll need to generate ad-hoc signatures to run the code you compile on your own computer. So that's a no from me.

      1. Jason Hindle Silver badge

        Re: From a developer perspective, it's going to be a very mixed bag.

        Probably not. Worst issue I had with Big Sur was that it didn't like executables downloaded via a web browser. That is easy enough to sort out:

        xattr -d com.apple.quarantine <name-of-executable>

        I would hope anything build on the box would run absolutely fine without issues. I do Ja little ava (so none of those binary executable Shenanigans) but would be happy to give helloworld.c a quick try (thirty years, but I think I can manage hell world in c) to prove either way.

    2. Anonymous Coward
      Anonymous Coward

      Re: From a developer perspective, it's going to be a very mixed bag.

      “for most of the users that don't need to run up virtual machines”

      i.e 99.999% of users. Most of which won’t even know what a fucking virtual machine is. Most PC laptops sold also get sold with 8GB.

      1. Jason Hindle Silver badge

        Re: From a developer perspective, it's going to be a very mixed bag.

        We're not exactly discussing this in a place for most people.

      2. Dan 55 Silver badge

        Re: From a developer perspective, it's going to be a very mixed bag.

        And the first four words of the title are...?

        1. Anonymous Coward
          Anonymous Coward

          Re: From a developer perspective, it's going to be a very mixed bag.

          Most developers do not run or need virtual machines either. They and containers are fetishised by a cult. Again I would also question if many developers even know what they are.

          1. Dan 55 Silver badge
            FAIL

            Re: From a developer perspective, it's going to be a very mixed bag.

            The reality distortion field now makes us believe that no true developer needs a VM.

  14. Anonymous Coward
    Anonymous Coward

    I think that the real question on everyone's mind right now is:

    Can it run Crysis Remastered ? (At full resolution with all the bells and whistles turned on). :)

    Sorry, I know that it will not, but the thought of it made me smile.

    1. Dave 126 Silver badge

      The only x86 is game I've read of benchmarked on the M1 Mac Mini is a Tomb Raider title. 60fps at lowish res, 40fps at higher settings.

      Vague, I know, but sits with the view of the M1 GPU as being 'better than any other integrated graphics, better than entry-level discrete graphics'.

      Never mind. My purely anecdotal take on things is that Mac users who play games tend to just buy a PlayStation to do so.

  15. Fruit and Nutcase Silver badge
    Black Helicopters

    Coincidentally...

    and unrelated but a tenuous link to elements of the headline. The Strong Arm of the Law are looking for some Apple kit and the gang that made off with it in a hijacking on the M1

    "M1 lorry hijack gang stole £5m of Apple products, say police"

    https://www.theguardian.com/uk-news/2020/nov/17/m1-lorry-hijack-gang-made-off-with-5m-of-apple-products-say-police

    1. Socks and Sandals

      Re: Coincidentally...

      “The Strong Arm of the Law”

      I see what you did there...

  16. timrowledge

    The only bit that matters is that Squeak Smalltalk works nicely on the M1 Macs as well as several other ARM64 devices like Pi.

  17. Joe Gurman

    Buehler? Anyone?

    Care to guess what percentage of Mac users use even one piece of open-source, user-installed software on their Macs? Or perhaps I should say, what fraction of one per cent?

    As you clutch your pearls and cry, "Think of the devs!" please remember that Apple provides several good ways of developing apps without third party libraries or code.

    1. Archivist

      Re: Buehler? Anyone?

      Microsoft Office, Libre/Open Office, VLC, Dropbox. Most people will have at least one of those and there is no advantage to using the App Store, even if they're available.

    2. Anonymous Coward
      Anonymous Coward

      Re: Buehler? Anyone?

      "Apple provides several good ways of developing apps without third party libraries or code."

      Erm... Apple uses LLVM, so no.

      Also the OS is BSD, and the command line is ZSH...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2021