I'm not following you. Apple owns both OSX and iPad OS. They wouldn't lose any money by putting both on an iPad. It's possible that it could be a "tight fit" for a lot of the current iPads, as you may want at least 8GB RAM and 256GB storage space for an enjoyable experience.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Bloomberg: Apple's First Arm Mac to Launch by 2021 With 12-Core Processor
- Thread starter MacRumors
- Start date
- Sort by reaction score
I have what may be a dumb question, or maybe it's here in the 30 pages, but would windows/linux/etc. still run in a virtual environment with ARM processors?
I don't disagree regarding making the iPadOS experience better when "docked" let's say, but knowing Apple over the years and the direction they are headed with iPadOS, I just can't see the point in offering something like that. Proper windowing, external display support, and maybe a docked vs undocked approach where elements adjust are probably the most we will see. Looking at it from a business lens, makes it even more less likely they would offer macOS on iPad.
They will lose money. You're right about having both, but in developing macOS for the iPad will not be free And to maintain two separate code bases will introduce a higher risk of bugs and instability. In short, they won't markedly improve their situation, now will this improve sales but yet there are significant downsides.
This was a post by a former Apple engineer on an Appleinsider thread (post #122). Agree or disagree?
"Thunderbolt 4.0 will leverage PCI-E 4.x and newer and no Apple won't be switching to USB 4.0 and jettisoning Thunderbolt. Thunderbolt is directly tied into PCI-E which allows all that bandwidth and low latency professionals require. USB 4.0 is offering 40Gbps, while Thunderbolt 4 offers 60Gbps.
However, in reality PCI-E 5.0 arrives with Zen 4 in one year and in 2021 that is most likely when Thunderbolt 4 arrives, thus Intel will be moving the goal posts again, but since they opened up Thunderbolt to a royalty free protocol AMD can walk in with their own Controller or third party controller with Thunderbolt 4 communicating with their new Infinity Architecture allowing a solution that Intel can not match but will be a me too in the end."
2022 Computing will change drastically and it won't be ARM.
"Apple is heavily investing into Machine Learning and the Data Center. They aren't going to be using ARM to do all the heavy lifting and Intel can't fill this void. Neither can Nvidia.
EPYC CPUs are followed rapidly by Ryzen CPUs and this is the year of Milan"
"July 2021 GENOA arrives and August it's cousin arrives for Consumers as Ryzen 5000 CPUs, followed by Ryzen 5000 APUs with RDNA 2.x for laptops.
NOTHING APPLE develops on ARM will ever compete on this scale and they know it. They are augmenting their ARM offerings like the new iPadOS and keyboard solution to folks who want just a bit more Laptop like options, but can leave the keyboard at home.
ZEN 4 will either default to 12 or 16 Cores and 24 or 32 threads [assuming they don't increase the number of threads each Core can produce per operation]. That means Threadripper will be 128 Core/256 Threads at the top end and 32 Core/64 threads at the low end/entry level.
12 ARM cores with 8 + 4 extremely low power will do what again? It will provide a new iPadOS optimized and slightly beefier solution giving the OS more multitasking features, but not much more."
Why Apple will move Macs to ARM, and what consumers get - General Discussion Discussions on AppleInsider Forums
There are obvious hardware benefits when Apple shifts to ARM processors, and also underlying software ones, which mean that the move is going to bring advantages both to the company and to us.forums.appleinsider.com
Agree. I wish more people looked at future trends and projections.
AMD has been executing the plan that was laid out three years ago. I would give them the benefit of the doubt.
I respect your opinion, but perhaps AMD has changed for the better since you were employed there?
Not the way you're accustomed to. ARM CPUs cannot virtualize x86 operating systems. If it does work it would require an emulation layer that would significantly reduce performance.
“It’s not impossible, it’s just not available” is not a super compelling defense for a move away from a Bootcamp capable computer. A lot of stuff current Mac users take for granted is going to be unavailable and remain that way.
Again, I'm not following. Isn't this entire thread ties to a story about them developing OSX for ARM? So, they're already doing the work. The iPad could have iPad OS *and* this new OSX for ARM and simply switch between the two modes when docked/undocked.
Actually you're wrong game exclusives are really not what they used to be - There are PlayStation exclusives that can now be played on the Xbox.
PS4 games shock as top PlayStation exclusives now playable on Xbox One
What I'm saying is there may very well be an ARM based laptop from Apple but it doesn't make sense for Apple to put macOS on the iPad
They'd lose Mac sales.
But actually, I wouldn't want MacOS on an iPad, but on an iPhone. Attach keyboard and monitor, use the phone screen as a trackpad, and you have a desktop computer. If I can attach a hard drive, that would be great. If they build ARM-based Macs, and suddenly any iPhone user can have a reasonably powerful MacOS desktop for not much over $100, that would be a blow to Windows.
I'd make MacOS vs. iPhone mode not dependent on whether it is plugged in, but either through a hardware switch, or the combination of plugged in + laying flat on the desk. So if I get a phone call I can just pick up the phone and take the call.
Throwing a mild derivative of the A14X at a one port wonder MacBook ARM upgrade doesn't really scale to covering the rest of the Mac line. That gets increasing misaligned as move up the product line. Apple has zero demonstrated capability in the upper end iMac BTO , iMac Pro , Mac Pro zone. That's isn't likely "before too long" a replacement.
Apple could choose to leave that space and just use ARM to cover the bottom "half" of the line up. It is also highly uncharacteristic of Apple's previous moves to lauch into covering several Mac products with 4-5 different variants of concurrent CPU packages ( which exists now in Mac product line up and does not exist in the in the iOS/iPadOS/etc product lines. ). The whole iPad Pro line tends to go to A__X when there is major upgrade. Same with iPhone. Same with AppleTV , etc. There are "last year's" products with older versions of the A-series in them but not multiple concurrent implementations. The only variation is the iPhone / iPadPro split (e.g., A14 and A14X ).
But that is a narrow pivot off the real volume driver. A14 (and "upside" for a A14X). This rumored extra upside of the A14X may not mean much in terms of differentiation if all it is doing is powering a one port MacBook wonder. ( which means constrained to the port constraints of the iPad Pro). From the Bloomberg article doesn't disagree with that path. So if that is path Apple is on then x86 won't be disappear soon because Apple isn't doing all the work necessary to make it completely disappear.
A workstation space optimized ARM chip going to appear from a 3rd party and Apple jumps on board ? I wouldn't bet on that. With Intel and AMD competiting heavily in that space over next 2-3 years there isn't going to be much of a easy open window to walk through there. ARM server targeted CPUs are a bad fit with macOS since they are primarily targeting high core counts. ( macOS doesn't deal with very high core counts well at all. It isn't likely Apple is going to hyper tune it for that either as it is entirely not needed for the shared kernel of the mobile products. If Apple dragging their feeting 6 years on a Mac Pro upgrade that highly unlikely a priority. )
Apple could put their desktop products in "Rip van Winkle" mode and ride it out with stale, trailing edge Intel products frozen in time for several years. That is not likely to turn out well for them if they try.
Apple's "pattern in past" don't match up to this current context. Apple got off of Motorola 68K because it was the 'end of the line' for the 68K in personal computers. Motorola was downshifting it into the embedded space and the 88K was the replacement in the personal computer / workstation space. That got traded to PowerPC. Apple was suppose to get two competing suppliers out of it (IBM and Motorola). But sticking with 68K wasn't even a viable option at all.
Apple's transition from PPC to x86 was to a more complete, faster, and robust line up with two major suppliers. Yes Apple picked Intel but AMD was there as a back-up foil. The PPC had also pragmatically run into a dead end. A healthy ecosystem had not build up in the personal computer space for the architecture so couldn't support multiple vendors building products almost solely for Apple. And Apple wasn't willing to pay to get the work down by themselves. So they bolted to a bigger ecosystem with shared costs, broader infrastructure , and less risk. Without Apple, PPC ( similar to 68K went for embedded market ) and reforked into Power 6-10 ( at costs not viable for Apple).
The current context x86 isn't dying. It isn't being "turned off" like 68K. It isn't dropping down to one vendor in the personal computer space ( AMD is far more highly competitive than in years past and Intel is down but not completely out. Neither one is getting out of the business and will each likely invest billions over next couple of years to stay in.)
As long as there a viable contractor(s) that is going a very good job then Apple consistantly outsources that work. Apple largely doesn't make their own displays. RAM . NAND . desktop GPU. Intel has been screwing up. If they were a RAM , NAND , or desktop GPU vendor with the track record they have put in over last 3 years, then they would have been tossed by Apple in 2020. There is a bunch of 'dot the i's , cross the t's' work they do coupled to the CPU+PCH+Thunderbolt . ( Also didn't help AMD much either when they were actively throwing roadblocks for Thunderbolt in years past. )
Apple could choose to "go it alone" and compete with AMD, Intel, Nvidia, Hwawei, Samsung, and Qualcomm all at the same time. Apple has some (not all ) talent and a hefty band account, but that is just asking for trouble. Going product "blow for blow" across wide product category is exactly what Apple doesn't do well. There is little sign that Apple is trying to get into the very broad spectrum CPU business. That isn't "short sighted" . Apple simply does not have a desktop class solutions. Nor do they have any reputable track record in that space. Their ARM desktop track record is AppleTV at this point.
[automerge]1587741570[/automerge]
There is linux for ARM. I used it on a desktop ten years ago, so that should be just fine today. There is Windows for ARM, but I never heard of anyone using it. I think having a virtual environment at all won't be the highest priority, but will come. You'd need an x86 emulator to run Windows; someone might decide to build that if there is money in it for them, Apple likely won't.
I'd assume that the low end machines would switch first. People using multiple VMs running Windows or Linux would want a very powerful Mac anyway, so the high end Macs will stay on sale longest. The only ones hurt are employees who need to run Windows, and who get a Mac just because it's fashionable and never run anything than Windows on it. (I've known a few of these. Lots of people needing to run Windows in a company, and the ones with a bit of political power got MacBooks, the others had to take what they were given - the cheapest Dell laptops). They won't be able to get a brand new MacBook Air rather soon.
Me and my wallet would love a MacBook with 256GB SSD and 2TB hard drive. Especially my wallet would love it.
Many years ago I knew lots of people with 68k based Atari computers. When PowerPC reached 150 MHz, some hackers made the 68k simulator on the PowerPC simulate a 68K Atari computer. And it was the fastest Atari computer ever.
Last edited:
Did you know that you can evaluate the function f(k) = a sin (kx) + b cos (kx) for consecutive values of k with just one floating point multiplication and addition? f(k+1) is a linear combination of f(k) and f(k-1), there is a constant dependent on x that you need to calculate once.
[automerge]1587743026[/automerge]
Fair play to AMD. They've gone up against the Intel and Nv' juggernauts and got back in the ring after being out cold on the canvas.
You have to respect that. They changed the value and performance equation whilst Intel were living the large of all complacent and dominant corporations in the sleep of premium profits.
Azrael.
[automerge]1587743181[/automerge]
...and as for gpus. They're done alright. Bringing mainstream performance and value back to the GPU market.
As for the high end. They're on their way...back care of the PS5 and the BiG Navi for PCs.
Azrael.
Fewer. ARM / x86 are much more similar than x86 / PowerPC. Capabilities are practically identical, and both are little-endian. Moving from PowerPC to x86 the biggest problem was code that checked if it was compiled for PowerPC, and assumed that if it wasn't PowerPC, it must be 68k... We don't have that now
[automerge]1587743300[/automerge]
Change a switch in the compiler, build, and you're done.
Disagree. So does Intel.
Intel Teases Thunderbolt 4, Light On Details
and
What Is Thunderbolt 4? Tiger Lake Tech Isn't Faster, Thunderbolt 3 With a New Name
Thunderbolt 4 appears to be mostly a re-branding of Thunderbolt 3.
Thunderbolt 4 is primarily matching up the USB 4. There is no PCI-e v4 element to it at all.
It is pretty doubtful Thunderbolt 4.1 ( or 5 ) ( and whatever USB4 increment. More likely a 4.1 than a 5.0 will be coming to USB any time soon) will get entangled with PCI-e v4. Thunderbolt 4 is more so going to be Intel "double checked and certified" the base USB4 implemenation that didn't leave out any of the almost entirely optional Thunderbolt parts. USB 4 only optionally includes Thunderbolt interoperability.
PCI-e v4 (and above ) have lead to tighter (shorter) distance constraints. Thunderbolt and USB need that like a anoter hole in the head. Coupling to USB highly likely means slower bandwidth evolution because the standards consensus problem is orders of magnitude bigger. ( far, far, far more 'chefs' in the kitchen who don't agree. )
PCI-e v5 is more so going to be a CPU-CPU and CPU-GPU memory bus interconnect ( e.g. CXL layered on top) than a random PC port I/O port issue. Relatively very short internal distances amost exclusively inside of a system.
Thunderbolt is not PCI-e. It isn't required to mimic ever move that PCI-e makes. It isn't "external PCI-e".
[automerge]1587744065[/automerge]
Not true. In fact, Apple went through significant gyrations to provision a USB-A output from the Intel PCH out to the the detachable I/O card on the Mac Pro 2019. ( also provisioned an internal USB-A socket inside the device. ) All the Mac desktop systems have USB-A. So de-supported? Not even hardly.
Mobile systems? Yes. For keyboards and Mourse? Yes. Across the whole Mac product line? No.
The Mac product space is not simply just the entry level laptop systems.
Last edited:
There is 0 chance that the mac cpu is a derivative of the mobile. It’s the other way around (despite bloomberg’s dopey reporting). I was a cpu designer for many years. You design the most for the most performant product, then scale downward. Not the reverse.
And the leaked numerical part number for the mac part indicates it is a major revision, changing to a 1 in the third digit for the first time.
The Bloomberg article tells us nothing about the form factor or ports on whatever the first couple of machines will be.
Plenty of technically superior processors (Alpha being one example) failed because of the factors like companies refusing to port software to the architecture (for economical and other reasons).
I see a societal problem here. Is it worth for humanity to support two fully developed ecosystems for desktop/server? Right now we have only one fully developed ecosystem (x86). Other ecosystems exist (Power, ARM) but they serve niche segments. To replicate everything that exists on x86 requires tremendous amount of resources. We started with plenty of ecosystems. For example, Windows NT supported IA-32, x86-64, ARM, DEC Alpha, Itanium, MIPS, and PowerPC. We had even more platforms before that. Eventually they all died because of the amount of resources required to support and develop fully fledged ecosystem.
Guys, I know it is possible that the new A series desktop Mac will have a line voltage plug and lack a battery UPS like AppleTV and HomePod and Mini do now, but I don’t understand how that is preferable to having a USB C connector for power, and a battery?
is there any evidence that Apple plans to run an ARM chip at 40wTDS with active cooling like desktop i7s?
is there any evidence that Apple plans to run an ARM chip at 40wTDS with active cooling like desktop i7s?
I think the de-support was referring to USB-A on MacBooks, they are all USB-C.
I reckon we will know for sure the Apple direction after the WWDC in June. The conversion to Apple CPU chips has been rumored as "next year" for quite a number of years now, but you have to give the developers some room to convert and re-deploy their programs for the newer chipsets, and that would probably take them a year or two, including either access to prototype hardware or a decent emulator (or both). If there is no announcement on direction at the 2020 WWDC, then that would mean 2022 earliest. Apple would also need a decent x86 emulator, so users can run the old stuff, that is non-trivial and I have not seen them move to buy one (IBM has the technology, as do a number of others, but there is some licensing stuff that you have to work through). Complicating things is that Intel and AMD are in a price+performance war, it is hard to be cost effective there in your development efforts, one could argue that the next mobile chipset switch for MacBooks should be to AMD. Apple has the low end covered, but not the higher end mobile chipsets, granted what it has would be faster with active cooling. Apple has been through the CPU chip conversion a couple of times, so they know this (granted, some, like Woz, are probably retired or semi-retired, so they may have forgotten some of it).
I reckon we will know for sure the Apple direction after the WWDC in June. The conversion to Apple CPU chips has been rumored as "next year" for quite a number of years now, but you have to give the developers some room to convert and re-deploy their programs for the newer chipsets, and that would probably take them a year or two, including either access to prototype hardware or a decent emulator (or both). If there is no announcement on direction at the 2020 WWDC, then that would mean 2022 earliest. Apple would also need a decent x86 emulator, so users can run the old stuff, that is non-trivial and I have not seen them move to buy one (IBM has the technology, as do a number of others, but there is some licensing stuff that you have to work through). Complicating things is that Intel and AMD are in a price+performance war, it is hard to be cost effective there in your development efforts, one could argue that the next mobile chipset switch for MacBooks should be to AMD. Apple has the low end covered, but not the higher end mobile chipsets, granted what it has would be faster with active cooling. Apple has been through the CPU chip conversion a couple of times, so they know this (granted, some, like Woz, are probably retired or semi-retired, so they may have forgotten some of it).
I don't think they're going to purposely *not* expand on the vision they're already evolving on with the iPad just so they can still sell traditional Macs. The iPad Pro + Magic Keyboard/case is a significant step towards making the iPad Pro into Apple's take on the Microsoft Surface. But when docked, the UX is lacking. I think they're already moving in the direction that I'm talking about, and it's just a matter of time.
Oh, I welcome that idea, too. In fact, I wrote about a similar concept over 10 years ago on a blog I used to run.
Yeah, the specifics for how it/you would switch from one OS to another will need to be figured out. If they go this route, I imagine they'll do a pretty good job of implementing a fairly elegant solution.
Short answer (as already given): it won't run x86 Windows or x86 Linux with anything like the efficiency of a x86 Mac. Potentially, they could run under an emulator/translator that simulates an x86 chip in software but that is likely to be horribly slow.
If you currently rely on this then you probably won't want to be queuing up for an ARM Macbook on day 1 (although you might want to "think different" over the next few years and look at alternatives).
More complex answer: in theory, it could run ARM Windows and ARM Linux efficiency - there are several ifs, buts and maybes to do with system architecture/firmware that would affect whether an ARM equivalent of BootCamp was feasible, and we can't know the answer to until Apple finally announces something. Also, it's up to MS whether they license ARM Windows for Mac hardware. Whether Parallels or VMWare Fusion would get ported (they're not going to be a case of 'tick the ARM box and re-compile!) with no market for x86 Windows is a good question - but MacOS has basic virtualisation support built in (as used by Docker for Mac) and ARM chips do support virtualisation - but AFAIK that doesn't include graphics access, so it would be mainly good for server/command line stuff.
At this stage, ARM Linux is pretty well developed with tons of compatible software and (say) Docker running on ARM Linux would be a very useful tool for web/server development (most of which doesn't give a wet slap what processor it is running on and at worst you'll only need an x86 for later-stage testing of any binaries you produce). ARM Windows, however, is probably not what you wanted Windows for, although it does have x86 translation/emulation for Win32 apps (64 bit promised) and while the reports on the performance of that are... underwhelming it should be much better than running x86 Windows under emulation (...because only the application code is being emulated whil all the operating system calls, UI rendering etc. are running natively).
However, Tim Cook isn't going to break into your house the day after WWDC and take away your Intel Macs. Apple have just launched the Mac Pro - which will probably be the last Mac to go ARM, if only because of all the obscure third-party plug ins that pro media types rely on - and if they don't support/update that for at least 3-4 years it will be nature's way of telling you never to rely on a single hardware vendor for anything important.
Looking long-term: MS seems serious about ARM Windows this time around - it is early days at the moment but it isn't dead yet, and a lot of modern Windows-specific software development is using the .NET framework which is processor-independent (not to mention the increasing importance of web-based and mobile development vs. traditional Windows apps). For non-developers, things like banking and tax apps that, previously, have tied people to windows are increasingly being replaced with web apps. Give it another few years, and Windows x86 binaries could be the sole domain of hardcore gamers and corporate computing - neither of which have ever been Apple strong points.
As for Linux - that's already mostly about servers/web development (most major Linux apps have MacOS versions anyway - it's not like you need a Linux VM to run LibreOffice or Inkscape) which can either be done in ARM Linux, a x86 Linux instance in the cloud or a cheap PC sitting in your basement. Yes, you'll need an x86 box to compile/test on if you're working on traditional Linux software - but, increasingly (with Amazon and others rolling out ARM-based servers and a zillion users with Raspberry Pis and Chromebooks), you'll need to compile/test on ARM Linux as well.
One of the strengths of Mac and MacOS is that it isn't pinned to the past by legacy code the way Windows is - it would be a shame to let Windows hold it back. Even if Windows compatibility was a selling point back in 2006, there have been huge changes in the computing world since then.
We have no idea what sort of power connection an ARM desktop mac will have. Changing the processor doesn’t really affect that.