Apple Transitioning Macs to Custom Chips With 'Incredible' Performance Starting Later This Year

[SteveW928]

Given that they are currently on macOS, and consider the Mac market to be important for them. Again, unlike Windows, Apple has made it clear that all their machines will be ARM, so if they want to stay, they will need to be on ARM. Again, for companies that are on Apple’s tool chains and use Metal, porting will be very easy.

Yeah, but will they continue to support Mac? My understanding (given I'm not a software dev) has been that it has been fairly easy to make Mac ports on x86 w/ OpenGL/CL, as most of these companies are Windows-centric. Will that still be the case once Apple moves to ARM/Metal?

I think you might be right here. They have a potentially large advantage that they do not have to pay anyone’s profit on their silicon, so they can build a competitive console at a lower cost, especially if they make 30% on games delivered to it.

I think we'd mainly see it on the low end, where Apple might use the savings to introduce low-end machines priced more like AppleTV/iPhone, and at the high end where Apple pays huge premiums to Intel (ie. Xeons). The question is whether they'll actually take advantage of that vs just having a higher profit margin.

 

[MikeZTM]

My point is that they didn’t.
Windows RT is not “the real Windows“. It never ran “real windows” applications.
The first full Windows release on ARM was Windows 10 in 2017.

It was real windows if you jailbreak it. And a lot of open source software was ported to it using preview version of Visual Studio 2012.
7zip, Filezilla, and much more.

And it have native win32 arm32 version of Office 2013 bundled with it.

 

[SteveW928]

AMD and Intel have built and shipped multi-threaded CPUs - Apple hasn't. AMD and Nvidia have shipped discreet GPUs - Apple hasn't. Intel will beat them to the punch also.

I am not saying that they couldn't build a 20-core/40 thread system - (although I would point out that they don't have any multi-threaded SoC at this point. AMD has developed 4 way SMT, that will end up in either Zen 4 or Zen 5. Most folks here wouldn't have any use for a 16 core/64 thread computer, but I certainly do).

They could - I just have no confidence in ability - Apple hasn't been able to walk and chew gum for a while now.

Fair enough... I didn't realize that. Good info. That said, I'm sure they *could* do it, but I'm a bit with you in whether they have all their ducks in a row, so to speak.

 

[sirozha]

Ah yeah, thats what can happen if there is no GPU/video driver for your device. Maybe it is also a memory problem on top, since 1fps is still slow even in the absence of a video driver. Do you have the 4Gbyte version of the PI4?

Thinking about this, i consider it not very likely that Apple will write video drivers for Windows. Of course if they are ambitious and want to have the best Windows Laptop on top of having the best MacBook they might - at least financially it would be a peanut investment for Apple.

Yes, 4 GB Raspberry Pi 4B

 

[mattferg]

The Surface Pro X runs an emulated x86 32 bit version of MS Office...why?

It doesn't, and hasn't for some time. It's a compiled hybrid portable executable. The code is native ARM code, but it uses x86 calling conventions so that x86 binary plugins are compatible. So while the front-end executable is x86, the libraries that it loads in the backend to do most of the work are native ARM64. Best of both worlds, no performance hit. Imagine the Mac solution will either be similar, or lose plugin support.

 

[Gerdi]

The Surface Pro X runs an emulated x86 32 bit version of MS Office...why?

I think i already explained the Office situation earlier. Microsoft had 2 options, either compile it for ARM and losing compatibility with 3rd party plugins or emulating Office and make it compatible with all 3rd party plugins. However they figured there could be a third option.

Short excursion on the topic:
An application consists of the executable itself (e.g. the *.exe), supplemental libraries (*.dll), system libraries provided by Microsof) (*.dll) and then the OS kernel services itself plus the kernel drivers.
If you start a 3rd party x86 program, Windows will use the x86 excutable, all DLLs as x86 versions and emulates all the containing code - when the app is doing a kernel or driver call, the execution will jump out of emulation an will execute driver and kernel code natively ARM. In this sense every application is only partially emulated and is running partially native.

So what Microsoft did is, they where compling special version of the DLLs as ARM, however with a special shim/wrapper, which makes the ABI compatible with x86, such that emulated x86 application can directly call into native ARM libraries. Microsoft is calling these CHPE (compiled hybrid portable executable) libraries.

Thats what they are doing with Office. Only application level (*.exe) is emulated but all libraries, kernel and driver calls are native ARM. This technique makes Office fly on the Surface Pro X as there is only very small emulation penalty and all the 3rd plugin compatibility is perserved.

 

[MacCheetah3]

My son will be attending school remotely in the fall, so I’m considering an ARM-based Mac for him for this fall (or winter). He didn’t have a computer until the COVID-19 pandemic hit, and when the schools closed, I gave him my old 2012 Mac Mini for his remote school lessons. Now he started streaming Xbox One NHL game from the downstairs Xbox One to the Mac Mini in his room by booting into Windows 10 on his Mac.

If I wanted to get him an ARM-based Mac to replace his current 2012 Mac Mini, would he be able to game on his computer in his room by streaming his NHL games from the Xbox One sitting in the living room? Somehow I doubt it. Mind you, this is not an enterprise requirement, but the ARM-based platform already presents a limitation in this simple use case scenario.

His school doesn’t require a Mac. Any computer works for the live online lessons, even ChromeOS works just fine. Guess what computer will serve all of his computing and gaming needs going forward? Not an ARM-based Mac. He will probably end up with a Dell XPS.

My point is this: transitioning to ARM is a good thing. Dropping Windows for X86 is a disaster - both for consumers and for businesses. But then again, Apple really couldn’t care less. Macs are but a small drop in their corporate profit bucket.

You addressed your own complaint... Choose the right tool for the job. If a Mac with a non-x86/x64 CPU won't accomplish your wants/needs, then don't get one.

Someone needs to design a box you plug into your thunderbolt port that has an x86 on it and lets you run software on it.

I mean technically that could just be an x86 Mac mini, using ip over tb, and software each side to copy the binary, setup a socket, etc over eg ssh. Heck for VMWare you can already use it to connect to a remote esxi server over the network.

Intel NUC (?)
I know it's not the solution you envisioned -- maybe someone should develop a more streamlined version?

 

[djjeff]

IMO, software/compatibility. x86 just isn't as important any longer (in the big picture, I personally need it).

The question isn't about now, it's about then. Why, back when Apple made the transition to PPC, did they not go with the superior ARM design and save Mac users the headache of two architecture transitions?

 

[Gerdi]

The question isn't about now, it's about then. Why, back when Apple made the transition to PPC, did they not go with the superior ARM design and save Mac users the headache of two architecture transitions?

Back in the days, Apple was for historic reasons in bed with Motorola. They both teamed up with IBM to develop the next generation architecture - which turned out to be PowerPC. Acorn/ARM - a small British design house - was not even on the radar.
They surely have hoped that the world outside Windows would jump on PowerPC bandwagon as successor to Motorola 68040. But as it turned out, the workstation manufactures all went with their own RISC architecutre be it PA-RISC, MIPS, SPARC or ALPHA. You could argue, that the workstation crowd killed themself with such a fragmentation.

 

[djjeff]

Back in the days, Apple was for historic reasons in bed with Motorola. They both teamed up with IBM to develop the next generation architecture - which turned out to be PowerPC. Acorn/ARM - a small British design house - was not even on the radar.

A small British design house with a superior architecture. The answer that you've been dancing around is that there is nothing superior about ARM.

 

[cmaier]

It does not. If it would, they would've shown more relevant infos and of course hard facts aka numbers.

It does. There is no computer yet, and numbers for a chip without a bus, memory, etc. are meaningless. And if they reveal the computer numbers, nobody would buy an intel machine in the mean time.

 

[Gerdi]

A small British design house with a superior architecture. The answer that you've been dancing around is that there is nothing superior about ARM.

Why do i have to list the same points again? Remember load-store-architecture, memory model etc.? And as i mentioned, I do not consider ARM ...especially the architecure revision back then superior to Alpha, as Alpha also had all the features i listed - it was just x86 which was fundamentally an architecture of the 70s.

 

[joeblough]

Back in the days, Apple was for historic reasons in bed with Motorola. They both teamed up with IBM to develop the next generation architecture - which turned out to be PowerPC. Acorn/ARM - a small British design house - was not even on the radar.
They surely have hoped that the world outside Windows would jump on PowerPC bandwagon as successor to Motorola 68040. But as it turned out, the workstation manufactures all went with their own RISC architecutre be it PA-RISC, MIPS, SPARC or ALPHA. You could argue, that the workstation crowd killed themself with such a fragmentation.

but that isn't true - apple collaborated with Acorn to create ARM-based processors for the newton. Acorn was definitely on apple's radar.

 

[djjeff]

Why do i have to list the same points again? Remember load-store-architecture, memory model etc.? And as i mentioned, I do not consider ARM ...especially the architecure revision back then superior to Alpha, as Alpha also had all the features i listed - it was just x86 which was fundamentally an architecture of the 70s.

Those points were insufficient to answer my question. Likewise I did not ask why Apple went with Alpha instead of ARM, I asked why they went with PPC instead of ARM. You keep dancing around this question.

 

[brinary001]

I assume that that is Linux compiled for ARM - not x86 instruction set virtualisation. Windows 10 (full desktop) with work applications (CadStar, Altium Designer, FEM simulators etc) does not exist for ARM - AFAIK.
Running x86 on ARM is going to take a huge performance hit.

Yeah, I think you're right about the demo. Well, I guess all we can do now is hope Microsoft builds up their ARM version of Windows enough that it has largely the same ecosystem as what they have now, but with the added benefit of allowing us Mac users to still virtualize it in some way like we're able to do now.

 

[dspdoc]

Tell me this isn’t Intel taking a major whack at Apple?! Holy smokes!! LMAO

 

[dgdosen]

SPARC, like ARM, is a specification and can be licensed. SPARC is RISC. MIPS was RISC. PPC was RISC. Alpha was RISC. If RISC is so superior why is it that most of them are gone in favor of x64?

Linus Torvalds had some thoughts about that:

RWT Forums - Real World Tech

content overridden

www.realworldtech.com

I wonder what he'll think if/when ARM 'home' computers take off.

 

[Gerdi]

Those points were insufficient to answer my question. Likewise I did not ask why Apple went with Alpha instead of ARM, I asked why they went with PPC instead of ARM. You keep dancing around this question.

I did gave a clear answer to that question. Why would you team up with ARM if you can team up with IBM and Motorola, your old long term allies? It was a bussiness decision in the first place. It was the idea, that the big three can move the industry and not some small company, whos was producing home computers for the british market.
I am also not sure if you are familiar with the situation back then, Acorn was a competitor to Apple with the Acorn BBC and Archimedes - they only later went IP provider as we know ARM today.

 

[Zdigital2015]

SPARC, like ARM, is a specification and can be licensed. SPARC is RISC. MIPS was RISC. PPC was RISC. Alpha was RISC. If RISC is so superior why is it that most of them are gone in favor of x64? SPARC is the only one on that list left and I don't even think Oracle is developing it further, only Fujitsu. One might argue PPC is still alive in IBMs POWER processors but those are not the same. What is it about ARM which makes it so superior to all of these other RISC processors which are, for the most part, gone?

Linux (and Windows NT). They became ubiquitous on x86 and could be run on low end hardware. It slowly, but consistently, eroded marketshare for Solaris, HP-UX, SGI IRIX, DG UNIX, et al. who refused to see they were going extinct and were still charging fat prices for their hardware and service contracts, which was being commoditized by “good enough” on Intel’s side. Alpha was killed by Intel through Compaq, which depended on x86 for their Enterprise business. Alpha was a monster, easily making Intel look like the crap it was and still is. MIPS, PowerPC and Alpha all ran Windows NT, but Intel basically paid MS to stop developing those versions to choke the market despite the fact that each ran circles around Intel.

Sun Microsystems killed themselves with their arrogance. SGI did too, even faster by not seeing the changes in GPUs that were coming, tried to embrace Windows NT and half assed it while trying to sell $100K visualization systems using strippers and booze. It did not end well for them. HP lost their souls to their printer market and the home PC market and said screw it to PA-RISC to save costs to engineer more printers. They’ve never recovered and are a shell of their former glory. So that’s why through a weird confluence of events by hook or by crook, that Intel finds itself in the cat bird seat, metaphorically speaking. Motorola died because of an incompetent CEO and IBM retreated from PCs to become a services and solutions company. PowerPC became abandonware along with OS/2 while POWER flourished selling fat margin AS/400s, DB2 and those sweet juicy service contracts.

Now, Apple has said enough is enough and now has the resources to build their own chip team, something they never had while they floundered through the early 90’s and basically had to join the AIM alliance as going Intel would have surely destroyed the company back then while Windows road high with 3.11 and 95. Got it? Good.
[automerge]1592954801[/automerge]

I did gave a clear answer to that question. Why would you team up with ARM if you can team up with IBM and Motorola, your old long term allies? It was a bussiness decision in the first place. It was the idea, that the big three can move the industry and not some small company, whos was producing home computers for the british market.
I am also not sure if you are familiar with the situation back then, Acorn was a competitor to Apple with the Acorn BBC and Archimedes - they only later went IP provider as we know ARM today.

Don’t forget the promises of Taligent and Kaleida that Apple, Motorola and IBM thought would decimate MS. Except for the idiot upper management at all three companies.

 

[joeblough]

Those points were insufficient to answer my question. Likewise I did not ask why Apple went with Alpha instead of ARM, I asked why they went with PPC instead of ARM. You keep dancing around this question.

the answer lies outside the purely technical -

1) they were able to rope in two other huge companies that thought maybe they could topple intel
2) ibm had a fab and libraries and layout engineers - partnering with acorn would not have brought these resources
3) motorola had designed several successful CPUs in the past
4) apple didn't necessarily know anything about CPU design at the time, making #2 and #3 important.

for ibm and microsoft, apple brought the sales volumes that made the project(s) viable

 

[tillsbury]

Tell me this isn’t Intel taking a major whack at Apple?! Holy smokes!! LMAO

Since when was "Intelbased" a word?

 

[Gerdi]

Don’t forget the promises of Taligent and Kaleida that Apple, Motorola and IBM thought would decimate MS. Except for the idiot upper management at all three companies.

Indeed

 

[djjeff]

Linus Torvalds had some thoughts about that:

RWT Forums - Real World Tech

content overridden

www.realworldtech.com

I wonder what he'll think if/when ARM 'home' computers take off.

What is your point?
[automerge]1592955458[/automerge]

I did gave a clear answer to that question. Why would you team up with ARM if you can team up with IBM and Motorola, your old long term allies? It was a bussiness decision in the first place. It was the idea, that the big three can move the industry and not some small company, whos was producing home computers for the british market.
I am also not sure if you are familiar with the situation back then, Acorn was a competitor to Apple with the Acorn BBC and Archimedes - they only later went IP provider as we know ARM today.

So it's your opinion they intentionally chose an inferior architecture?
[automerge]1592955507[/automerge]

the answer lies outside the purely technical -

1) they were able to rope in two other huge companies that thought maybe they could topple intel
2) ibm had a fab and libraries and layout engineers - partnering with acorn would not have brought these resources
3) motorola had designed several successful CPUs in the past
4) apple didn't necessarily know anything about CPU design at the time, making #2 and #3 important.

for ibm and microsoft, apple brought the sales volumes that made the project(s) viable

IOW they didn't believe in the superiority of ARM.
[automerge]1592955602[/automerge]

Linux (and Windows NT). They became ubiquitous on x86 and could be run on low end hardware. It slowly, but consistently, eroded marketshare for Solaris, HP-UX, SGI IRIX, DG UNIX, et al. who refused to see they were going extinct and were still charging fat prices for their hardware and service contracts, which was being commoditized by “good enough” on Intel’s side. Alpha was killed by Intel through Compaq, which depended on x86 for their Enterprise business. Alpha was a monster, easily making Intel look like the crap it was and still is. MIPS, PowerPC and Alpha all ran Windows NT, but Intel basically paid MS to stop developing those versions to choke the market despite the fact that each ran circles around Intel.

Sun Microsystems killed themselves with their arrogance. SGI did too, even faster by not seeing the changes in GPUs that were coming, tried to embrace Windows NT and half assed it while trying to sell $100K visualization systems using strippers and booze. It did not end well for them. HP lost their souls to their printer market and the home PC market and said screw it to PA-RISC to save costs to engineer more printers. They’ve never recovered and are a shell of their former glory. So that’s why through a weird confluence of events by hook or by crook, that Intel finds itself in the cat bird seat, metaphorically speaking. Motorola died because of an incompetent CEO and IBM retreated from PCs to become a services and solutions company. PowerPC became abandonware along with OS/2 while POWER flourished selling fat margin AS/400s, DB2 and those sweet juicy service contracts.

Now, Apple has said enough is enough and now has the resources to build their own chip team, something they never had while they floundered through the early 90’s and basically had to join the AIM alliance as going Intel would have surely destroyed the company back then while Windows road high with 3.11 and 95. Got it? Good.

Great write up but all irrelevant if RISC were the superior processor "architecture" we're led to believe by the Apple desktop ARM cheerleaders. Justify it any way you want x86 was competitive and, in some instances, faster than RISC. Therefore continued development in alternative processors wasn't warranted as it was in the early days of computers.

 

[Gerdi]

So it's your opinion they intentionally chose an inferior architecture?

*sigh* They did not chose an architecture at all. They decided to develop an architecture - and of course the plan was to develop the best architecture in the world.

 

[Zdigital2015]

The question isn't about now, it's about then. Why, back when Apple made the transition to PPC, did they not go with the superior ARM design and save Mac users the headache of two architecture transitions?

Asked and answered by myself and multiple other people in this thread.

It doesn’t matter what happened in the past anyways, that cannot be changed. Your fixating on irrelevancies. Chalk it up to different management back then. Your questions are pointless and tedious, accept or don’t, it doesn’t change anything. Unless you have a time machine we don’t know about.