Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

djjeff

macrumors 6502
Jun 10, 2020
318
162
I'll bet you 10 to 1 that Apple has something similar in the works, where it takes a macOS API calls and translates it into ARM calls on the fly, and 99% (number pulled out of my ass, yo) of the current applications would run without knowing what happened.
I have no doubt Apple will provide the ability to run x86 code on ARM. After all they did so in the past with the 68K to PPC transition as well as the PPC to x86 transition. The solutions for both were generally seamless, trouble-free, and went very smoothly. The only downside tended to be a loss of performance due to the emulation. I would be absolutely stunned if they did not do something similar for ARM.

These solutions worked very well as, at the time they were done, Mac users typically used Mac applications. Today there's an added element in that a Macintosh can natively run x86 software which means it can natively run Windows. The switch to ARM will negate that benefit. Despite the existence of Windows for ARM it appears users have little interest of moving to it.

IMO there are a couple of reasons for this:
  1. The performance gain, if any, offered by ARM is insufficient.
  2. The lack of native ARM Windows applications.
Combine these and users see no compelling reason to make a wholesale switch to a new architecture.

So what does this have to do with the Macintosh transition to ARM? IMO a significant number of people use Windows on their Macintosh. Be it natively (i.e. Bootcamp) or through virtualization (Parallels). I suspect they can continue to do so with ARM based systems but at what, if any, performance hit? It's unlikely the majority of Windows software will ever migrate onto another architecture due to overwhelming critical mass.

Thus the move from x86 to ARM may make a lot of sense for pure Macintosh uses but it becomes murky when it comes to cross platform (as in OS platform) use. IMO one of the strengths of moving the Mac to x86 was that one could buy a Mac and, if some of the software they needed wasn't available on macOS, they could use it under an alternative operating system (typically Windows) without penalty. Unless their ARM processor can reasonably emulate x64 this advantage appears to be going away.
[automerge]1591971988[/automerge]
To me 4 years going into 5 years is a longer delay than temporary. If they were building something for the government that cost upwards of 10 billion a year and it ran over by 4 years there would be congressional hearings to determine why.

And it's possible they will abandon 10nm entirely, move to TSMC or Samsung's 7nm nodes etc
I agree that it is a long time but, unless you intend to make the argument they'll never move off of 14nm, it is temporary. I agree it probably makes more sense to skip it for something smaller.
 
Last edited:
  • Like
Reactions: hobowankenobi

Flint Ironstag

macrumors 65816
Dec 1, 2013
1,334
744
Houston, TX USA
Thoughts about how an ARM based Mac Pro would affect the release of new types of MPX modules for the 2019 7,1...?
I would like to see a Mac (xMac) released in this ARM format with 1 MPX bay. There’s a huge market of wanna be Mac Pro users, and a lower priced model would support development of more MPX modules. They should be backwards compatible with 7,1 and ‘just work’.
 

djjeff

macrumors 6502
Jun 10, 2020
318
162
The dark grey is the A13 running in the current iPhones. The blue at the bottom is the Core 9900K (Skylake) and the orange is the Ryzen 3900X (Zen2). There is plenty of headroom to tweak things if Apple isn't constrained by the phone size (thermals and power) and small chip size (98mm2 A13 versus 180mm2 for 9900K and 80mm2 per CCX plus uncore stuff in the 3900X).
IMO there's too much difference between these devices to really draw any meaningful conclusion. I think the only thing we can conclusively say is that Apple feels making this transition is in their and their users best interests.
 
  • Like
Reactions: Adult80HD

Quu

macrumors 68040
Apr 2, 2007
3,442
6,876
I agree that it is a long time but, unless you intend to make the argument they'll never move off of 14nm, it is temporary. I agree it probably makes more sense to skip it for something smaller.

Sure sure, this isn't a permanent situation that they'll be on 14nm. They will make CPU's on a smaller process node whether their own or one of their competitors.
 

Waragainstsleep

macrumors 6502a
Oct 15, 2003
612
221
UK
They won't be able to provide an Intel emulator, if that's what your'e hinting at. ARM processors aren't powerful enough for that, as you can see with Microsoft's Surface X. As a matter of fact, I doubt ARM processors are even able to run a full version of MacOS properly. Apple with likely show a new version of iPad OS for larger laptops. The move to ARM will likely signal the end of MacOS
[automerge]1591872252[/automerge]

No I mean they provide tools for developer to recompile their software for the new architecture quickly and easily like they did with the Intel transition. Only even better.
You clearly don't realise how good Apple's A series chips are. They are matching or outperforming Intel chips at lower clock speeds and much lower thermal output. If Apple can put together a desktop spec ARM chip using their tech, they will obliterate Intel's offerings.
 
  • Haha
Reactions: chikorita157

high heaven

Suspended
Dec 7, 2017
522
232
I would like to see a Mac (xMac) released in this ARM format with 1 MPX bay. There’s a huge market of wanna be Mac Pro users, and a lower priced model would support development of more MPX modules. They should be backwards compatible with 7,1 and ‘just work’.

Agree. Apple is not making any normal desktop around $3000 which is a huge problem since Mac Pro used to be $3000 for a long time.
 

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
No I mean they provide tools for developer to recompile their software for the new architecture quickly and easily like they did with the Intel transition. Only even better.

And this will be way easier than the Intel transition because there are no endianness issues.

For a lot of apps, developers will upgrade Xcode, and Xcode will automatically start building an ARM version at the same time as the Intel version. The only thing that stands out as a possible issue to me is if Apple drops OpenGL completely on ARM.
 

zephonic

macrumors 65816
Feb 7, 2011
1,314
709
greater L.A. area
How sure are we that it will be ARM? I know Gurman’s article mentions that sideways but not explicitly. More like “Apple is gonna make their own Mac processors based on their i-device processors which happen to be ARM.”

It seems like the path of least resistance, but is there a chance they’d go a completely different route, or even make their own x86?
 

Quu

macrumors 68040
Apr 2, 2007
3,442
6,876
How sure are we that it will be ARM? I know Gurman’s article mentions that sideways but not explicitly. More like “Apple is gonna make their own Mac processors based on their i-device processors which happen to be ARM.”

It seems like the path of least resistance, but is there a chance they’d go a completely different route, or even make their own x86?

100% it will be ARM. No question at all in my mind. They can still add custom extensions to provide customised functionality, they're not restricted by their choice to go ARM it's just the foundation of the building they can build whatever they need on top of that.
 

DeepIn2U

macrumors G5
May 30, 2002
13,051
6,985
Toronto, Ontario, Canada
The last Power Mac G5 A1117, PowerMac11,2, was released October 19, 2005.
It was discontinued after the release of the first Mac Pro in August 2006.


The Power Mac G5 became "vintage" per Apple's rules five years after final sale -- August 2011

Much closer to 6yrs - just 2mths shy of 6yrs, not 5yrs. Still it’s 1 yr shorter than Apple typically lists a machine vintage at 7yrs. Somewhat a concern. The OS support is more critical as Apple typically upgrades their own software to the latest OS very quickly!
 

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
Much closer to 6yrs - just 2mths shy of 6yrs, not 5yrs. Still it’s 1 yr shorter than Apple typically lists a machine vintage at 7yrs. Somewhat a concern. The OS support is more critical as Apple typically upgrades their own software to the latest OS very quickly!

I think Intel Macs will get OS upgrades longer than PPC Macs did. A lot of the tools Apple makes now are cross platform with Windows and Linux (like LLVM/Clang and Swift.) Apple is going to be supporting stuff like x86 Swift basically forever, even if they aren’t making x86 Macs. So if they have to keep all the x86 tools going, might as well keep up x86 updates.

Very different than the PowerPC era where they were the only ones keeping the PowerPC tooling going, and it was only for themselves.
 

djjeff

macrumors 6502
Jun 10, 2020
318
162
You clearly don't realise how good Apple's A series chips are. They are matching or outperforming Intel chips at lower clock speeds and much lower thermal output. If Apple can put together a desktop spec ARM chip using their tech, they will obliterate Intel's offerings.
Perhaps because we haven't seen a desktop ARM processor from Apple? I am amused at how people continue to push ARM as some form of panacea for a replacement desktop processor. I keep reading about how good such a chip should be but we have no actual chip upon which to make such a determination.
 
  • Like
Reactions: ssgbryan

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
Perhaps because we haven't seen a desktop ARM processor from Apple? I am amused at how people continue to push ARM as some form of panacea for a replacement desktop processor. I keep reading about how good such a chip should be but we have no actual chip upon which to make such a determination.

Has nothing to do with ARM itself. Apple can build CPUs at 5 nm. Intel builds CPUs at 14 nm. Apple CPUs win.

If Apple had an x86 license they could build and design 5 nm x86 CPUs. But they don’t have an x86 license and they have existing ARM CPU designs so ARM it is.
 

Mauro Marcenaro

macrumors newbie
Apr 16, 2020
22
5
APPLE IS NOT A CORPORATE FOR PERFORMANCE BUT FOR SHARE HOLDER
arm apple will be a nitemare , video sound , photo, and rendering + plugins
are harded coded by hoise software and apple tell in few trick all will be ok ??

hahhah

this is an apple paid site,
we can see which posts, all is good all is great, but if apple has been selling broken keyboards for 4 years, ram at the cost of 24ct gold, from late 2013 until 2019, the benchmark cpu have increased by 70 -90%,
discrete gpu are 20 or 30 dollar chips, which are paid 500 dollars,
apple sells fairy tales, has a container of people who love fairy tales and pay for fabulous products
[automerge]1591996540[/automerge]
Are you saying TSMC can't build a desktop CPU at 5 nm? Cause if so AMD would probably like to know. They're using TSMC at 7 nm right now with an eye on 5 nm.
amd will eat apple chip, apple is looking for money and not performance, arm cpu will be good in 2025 , can be, but who buy apple first device with arm will lose time and money
 

djjeff

macrumors 6502
Jun 10, 2020
318
162
Are you saying TSMC can't build a desktop CPU at 5 nm? Cause if so AMD would probably like to know. They're using TSMC at 7 nm right now with an eye on 5 nm.
I am asking you which desktop ARM CPU you're referring to.
 

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
I am asking you which desktop ARM CPU you're referring to.

And I don't understand the relevance of the question. I'm talking engines and you're asking what language the driver speaks. It's not usually relevant. The first ARM CPUs were desktop CPUs, so I'm not even sure what you're implying. They weren't designed for mobile devices. It just had such good efficiency that it was easy to scale down to those devices.
 
Last edited:

djjeff

macrumors 6502
Jun 10, 2020
318
162
And I don't understand the relevance of the question. I'm talking engines and you're asking what language the driver speaks. It's not usually relevant. The first ARM CPUs were desktop CPUs, so I'm not even sure what you're implying. They weren't designed for mobile devices. It just had such good efficiency that it was easy to scale down to those devices.
Perhaps I can shed some light on my question. In the mid 1990's I used to participate in the comp.sys.mac.advocacy forums. Recently, due to nostalgia, I have revisited the newsgroup reading posts from that era. All kinds of predictions and speculation, from both sides, that, with hindsight, ended up being completely wrong. A lot of predictions based on announced but unreleased products. People claiming "victory" based on such products.

The same is repeating itself today. People implying ARM is the panacea of processor development. Based on the lessons learned from the 1990's I will no longer accept any argument based on unreleased products. If you can point to a desktop ARM CPU which supports your claims then please point me to it. Otherwise you're engaging in speculation which, IMO, is essentially worthless.
 
  • Like
Reactions: dspdoc

vbctv

macrumors 6502a
Sep 25, 2013
886
687
Cleveland, OH
I just can’t imagine an ARM CPU replacing a Xeon with any adequacy.

I’d bet Mini, MacBook(maybe a new MacBook Air), and the lowest-end iMac get Apple CPU’s while higher end iMacs, maybe a Mini-Pro, MacBook Pro and Mac Pro keep Intel.

I agree with this sort of. I say first device with ARM will be revival of 12 inch MacBook, then a budget Mac mini. Eventually MacBook Air & iMac. I think the Pro products including current Mac mini (which is basically a Pro model), Mac Pro, iMac Pro & MacBook Pro all stay on Intel for the foreseeable future...
 
  • Like
Reactions: dspdoc

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
Perhaps I can shed some light on my question. In the mid 1990's I used to participate in the comp.sys.mac.advocacy forums. Recently, due to nostalgia, I have revisited the newsgroup reading posts from that era. All kinds of predictions and speculation, from both sides, that, with hindsight, ended up being completely wrong. A lot of predictions based on announced but unreleased products. People claiming "victory" based on such products.

The same is repeating itself today. People implying ARM is the panacea of processor development. Based on the lessons learned from the 1990's I will no longer accept any argument based on unreleased products. If you can point to a desktop ARM CPU which supports your claims then please point me to it. Otherwise you're engaging in speculation which, IMO, is essentially worthless.

ARM is irrelevant. Apple can ship at 5 nm. Intel ships at 14 nm. Apple could be switching to MIPS or PowerPC and it wouldn't matter. It doesn't even matter if ARM is somehow worse than x86.

The only thing that matters is that Apple can produce processors at 5 nm and Intel isn't anywhere close. That will destroy Intel on performance regardless of anything else.

Any talk about ARM the instruction set is irrelevant if Intel can't produce processors on a modern process. At 5 nm Apple could build Mac Pros with laptop CPUs and it would blow away a Xeon. Intel can't make up that gap.

If Intel had a 5 nm process they could probably beat Apple CPUs. But they don't. So they can't.

AMD is this in action. I don't think AMD is better at designing CPUs than Intel. But they're at 7 nm and Intel is at 14 nm and they're steamrolling Intel. It doesn't matter how good Intel is at x86. It's irrelevant.

The performance speculation has absolutely nothing to do with ARM and everything to do with Apple and TSMC having 5 nm production.
 
Last edited:

djjeff

macrumors 6502
Jun 10, 2020
318
162
The only thing that matters is that Apple can produce processors at 5 nm and Intel isn't anywhere close. That will destroy Intel on performance regardless of anything else.
Then please point me to this desktop processor that Apple is producing that destroys x64 (I assume that's what you mean with your reference to Intel). I'm not close minded, I just want to see actual shipping hardware, not some advocacy speculation (which is what I saw a lot of in the CSMA newsgroups and which ultimately turned out to be untrue).

Enough with arguments presented as facts unless there are facts to support them.
 
  • Like
Reactions: ssgbryan
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.