Intel Expecting Apple to Transition to Custom ARM-Based Chips Starting in 2020

[bollman]

What is being neglected here is that the quality of programs (I refuse to call them apps) on the mac platform will decline. Most developers will be too lazy to make programs that work perfectly on the mac, they will just call "good enough" on the mac and therefore, there will be no need for the mac. There will be a massive increase in programs, yes, but they quality will be sub-standard.
For those who want to feel what problems you're facing when running mobile "apps" on desktop/laptops, I recommend trying Android X86 in a virtual machine like VirtualBox.
It works pretty well, but you can instantly tell that the UI that works on mobile is less than great, or even downright crap on a laptop.

Also, bootcamp will be lost, and virtualisation will be a lot less useful. Sure, bootcamp into vaporware win10 ARM, but nobody cares about windows per se, it's the programs that are key. And, I don't see it ARM catching on on the windows side of things anytime soon which means programs on windows will be x86 for a forseeable future and thus will run like stuck in mud in a emulator on ARM.
We who used macs before the x86-transition remember the horrors of running Connectix VirtualPC and trying to get _anything_ done in a virtual windows 98 running on MacOS 8.5 on a PowerBook G3 Lombard. Oh the humanity...

 

[cmaier]

I would argue that you are making assumptions that are only possibilities of the world outside the US, and my experiences differ from the leap of faith you are making. You are using American logic-which I get. I just don't think the global market is going to be as excited as you seem to be-and I do my best to not make assumptions; maybe I'm wrong and my reservations are unfounded, but your claims of definitive success are at best you being hopeful.

That's my biggest point.

Best luck-I'm just hoping the Pro lines stay Intel as long as possible.

It’s quite possible that the percentages are different in other countries, but they would have to be INCREDIBLY different to overcome the sheer volumes of 900million vs 80million. Remember - if iphone users in other countries wont buy macs because they don’t make monetary sense, then windows users in those countries also likely wont buy macs just to use boot camp. You can’t on the one hand say “well, of those 900million, 700 million are outside of the US, and none of them would spend money to get a mac” and on the other hand say “but they have so much disposable income that rather than buying a $100 box to run windows they’ll fork over a thousand bucks to run boot camp on a mac.”

And it’s not like people who buy iPhones can’t afford macs. There are macs that are cheaper than iPhones, after all. And post-arm there will likely be even cheaper options.
[doublepost=1550818464][/doublepost]

What is being neglected here is that the quality of programs (I refuse to call them apps) on the mac platform will decline. Most developers will be too lazy to make programs that work perfectly on the mac, they will just call "good enough" on the mac and therefore, there will be no need for the mac. There will be a massive increase in programs, yes, but they quality will be sub-standard.
For those who want to feel what problems you're facing when running mobile "apps" on desktop/laptops, I recommend trying Android X86 in a virtual machine like VirtualBox.
It works pretty well, but you can instantly tell that the UI that works on mobile is less than great, or even downright crap on a laptop.

Also, bootcamp will be lost, and virtualisation will be a lot less useful. Sure, bootcamp into vaporware win10 ARM, but nobody cares about windows per se, it's the programs that are key. And, I don't see it ARM catching on on the windows side of things anytime soon which means programs on windows will be x86 for a forseeable future and thus will run like stuck in mud in a emulator on ARM.
We who used macs before the x86-transition remember the horrors of running Connectix VirtualPC and trying to get _anything_ done in a virtual windows 98 running on MacOS 8.5 on a PowerBook G3 Lombard. Oh the humanity...

Some will decline. But a lot of software is not available at all, or is only available as an electron app. Having iOS-like versions of that software will be better than not having it at all, or having crappy electron apps.

 

[skitidetdu]

Since I work in creative fields I wonder what will happen with applications like ProTools, Cubase, the Adobe suite etc. Will they make A serious effort to port their products (again?) or is windows becoming an even more valid option?

 

[Reindeer_Games]

You can’t on the one hand say “well, of those 900million, 700 million are outside of the US, and none of them would spend money to get a mac” and on the other hand say “but they have so much disposable income that rather than buying a $100 box to run windows they’ll fork over a thousand bucks to run boot camp on a mac.”

I didn't-you just did. American consumers are more likely to buy more frequently than global markets. You are the one thats been focused on mass produced netbooks trying to access a bigger global marketshare-I'm telling Apple to serve their Pro users, Where-ever the live.

 

[cmaier]

Since I work in creative fields I wonder what will happen with applications like ProTools, Cubase, the Adobe suite etc. Will they make A serious effort to port their products (again?) or is windows becoming an even more valid option?

Windows is always a valid option. I think adobe will port - they are already working on a serious iOS version of many of their apps, after all, and they know they need to clear out the cruft and start with a clean architecture. Some of these apps wont be ported, but the hope is that many more people will buy macs, that developers can easily target the same code base to work well with both macs and the larger installed base of iPads, and so porting becomes more valuable to the developers.
[doublepost=1550818949][/doublepost]

I didn't-you just did. American consumers are more likely to buy more frequently than global markets. You are the one thats been focused on mass produced netbooks trying to access a bigger global marketshare-I'm telling Apple to serve their Pro users, Where-ever the live.

Ok, Americans buy more frequently. There are 200million active iPhones in america. There are 25 million macs. What’s your point?

When did i “focus on mass produced netbooks.” I never mentioned netbooks even once. Why should apple serve their pro users, who don’t make them nearly as much money as iphone users?

 

[ROGmaster]

Apple’s latest A12 chip is already blazingly fast without any cooling fans and a low TDP. Imagine the A12 in a MacBook Pro 13” with fans and a higher TDP. It’s gonna be a screamer.

The increase in performance will most likely be negligible, Geekbench already shows us A12's peak performance.
The A12X is not a SOC designed for higher TDPs, so again increasing the TDP will have small effects.

 

[cmaier]

The increase in performance will most likely be negligible, Geekbench already shows us A12's peak performance.
The A12X is not a SOC designed for higher TDPs, so again increasing the TDP will have small effects.

What does “not designed for higher TDPs” mean? When I designed microprocessors, we designed them to hit a certain performance at a certain TDP. But if they were put into boxes that allowed higher TDP, and we turned up the frequency and voltage, they always ran faster. Saturation might occur under certain bandwidth-limited workloads, but that’s rare. (You can’t speed up the external bus).

 

[kwikdeth]

folks, seriously, this is Apple we are talking about.
there are dozens, no hundreds of cases where its simply been a case of, "Deal with it."

Remember that PPC got one single OSX update after the Intel switch was announced. One.

You will have to replace everything. Again.

 

[cmaier]

folks, seriously, this is Apple we are talking about.
there are dozens, no hundreds of cases where its simply been a case of, "Deal with it."

Remember that PPC got one single OSX update after the Intel switch was announced. One.

You will have to replace everything. Again.

Well, almost everything. You can keep your sense of whimsy and your office 365 subscription.

 

[kwikdeth]

As long as ARM Macs don't require you to install off the App Store, I'll be delighted.

Never. T2 + ARM-based macs means complete and total lockdown of the MacOS, just like iOS. If its not on the App store, its not gonna happen without a jailbreak.

Also means the hackintosh community will basically be obliterated once they drop x86.

Apple wins all around. Shareholders rejoice. And if you complain, the locusts hordes of fanboys will say the problem is you.

 

[Reindeer_Games]

Why should apple serve their pro users, who don’t make them nearly as much money as iphone users?

Probably because they are still trying to sell several hundred thousand multiple thousand dollar machines that they have in stock. I love how my words are described as bold when yours are just venomous and cruel at times-I can take it, but it's mind boggling to say the least.

 

[kwikdeth]

Well, almost everything. You can keep your sense of whimsy and your office 365 subscription.

oh my god, thanks... nearly spit out my drink reading that one.

 

[0924487]

The mention of the T2 chip does not fill me with optimism for this move forward...

T2 issues are BridgeOS related. The transition to ARM based chip will run a derivative of iOS natively, since no x86 instruction set will be supported, just-in-time cross-assembling will not be needed, so BridgeOS will not be used for most tasks. BridgeOS will run as a module of the iOS derivative OS and Apple will likely use an low cost and low power intel Core M i3 processor that runs X86 macOS to support AppKit APIs and apps that must use x86 instruction set in early years of the transition to make users not totally pissed off. This also bypasses the need to beg Intel to license the x86 instruction set to them, which Intel may give **** to Apple in retaliation of its departure. So there will be two Darwin Unix Kernals running in parallel, one in x86 on Intel Core M i3 coprocessor, one on Apple ARM (very likely named E1/X1 processors). Even when BridgeOS crashes main OS will not be affected.

 

[aaronhead14]

For low end machines, I can see this working. But I don't see Apple switching to ARM for desktop workstations. They've got to stay with Intel for the high end users.

 

[ROGmaster]

Forget Windows games.

Windows Games sell way more computers than Apple's software.
But yeah, "forget windows games".

 

[cube]

Windows Games sell way more computers than Apple's software.
But yeah, "forget windows games".

Another one that does not understand what I wrote.

 

[YaBe]

I am not sure the Mac can survive this transition, or may be it's their way to bid farewell to the Mac and push the iOs thing...

Ah, sweet déjà vu… Remember the days of waiting for the magical G5… and then Intel was going to solve that dilemma forever and ever and ever?

I agree with you, but today the problem is hardly the fact that there are no new CPU avaiable, it's the fact that Apple is slow updating its lineup...i mean do we have any Mac running the 9 gen? They have been out for a while.

 

[bluecoast]

Makes sense. I believe that ARM’s roadmap focusses on an architecture optimised for multi-threading in 2019-2020, so this timeframe figures

As does the report that initially, ARM will only be at the low end. For that, I’d read the MacBook and iMac. I suspect that the Air will hang around for a while for those that want an entry level intel MacBook.

The mini is an interesting case as Apple seem to be positioning this as a dev box, so you’d expect a dev version of a Mac ARM computer to come out on the mini first. WWDC 2020 with the first wave of ARM Macs coming out soon after?

 

[ROGmaster]

What does “not designed for higher TDPs” mean? When I designed microprocessors, we designed them to hit a certain performance at a certain TDP. But if they were put into boxes that allowed higher TDP, and we turned up the frequency and voltage, they always ran faster. Saturation might occur under certain bandwidth-limited workloads, but that’s rare. (You can’t speed up the external bus).

Yeah things like 20% higher frequency and 7% better performance on average. That's what I'm talking about.
There's also other things to consider, a lot of designs when increasing the frequency and voltage they start to burn power like there's no tomorrow, things like 10% higher frequency, 22% higher power consumption.

 

[Reindeer_Games]

Another one that does not understand what I wrote.

I think I get it, but AAA title traditionally start on Windows and are ported. Thats the workflow most know and understand-I don't think I've ever seen a AAA title start in macOS and go vice-versa, but I do understand that Metal is different in a good way which may have been what I missed in your statement earlier.

Forgive us, but the statement is rather vague.

 

[Exponent]

Well, for a moment it was...

And there was the Exponential x704. I was the floating point designer on that., It was pretty fast and ran at around 2x the clock of intel at the time.

Shorthand: "x704" = "Welcome to signal integrity, before any good tools are available!" (In other words, an educational experience... A painful one...)

Regarding ARM Macs, the one area I know that will be a pain is _demoing_ (not actual real-world design using) EDA tools. Having a RedHat VM image running your software on a MacBook Pro is standard operating kit for a lot of CAD companies, semiconductor or otherwise. That's another pro market Apple will shut themselves out of.

 

[ksec]

I didn't bother reading the 21 pages of Comment. But two things.

1. ARM has plans to enter high performance CPU design with a very strong roadmap. ( https://www.anandtech.com/show/13959/arm-announces-neoverse-n1-platform )

2. What will happen to this year's Mac Pro? Assuming it is still coming this year.

 

[bluecoast]

So this is how Mac will die.
To all saying cool and great arm gpus are fast and maybe even better that iris pro etc. well blame apple and their greed for not putting a dedicated gpu on the Macs.
As for cpu power don't make me laugh arm going against i7,...... not even against an AMD.

I think that the writing on the wall has been for some time that Apple expects the Mac to be the ‘truck’ to the iPad’s ‘Car’ (as Steve Jobs once famously said).

I guess that they know that there will always be some people/businesses who prefer/need the pc form factor/functionality/power so I don’t think that they’re going to stop making Macs any time soon.

This way leaves them the space to add more functionality to iOS when it suits them rather than having to stuff it full of mis-matched features as that’s the only game that they have (see: Windows 10). The Mac can remain the Mac (with an injection of Marzipan and the same architecture so that it’s near painless for developers to support the Mac).

As to where the Mac will be in 2025... With things like VR and AVR I think that Apple will always need a powerful and flexible dev platform.

I’m sure they expect clamshells to gradually fade. This way if they have the same architecture across all of their products it’s not so painful for them to keep supporting the Mac for consumers for longer than they would’ve if they had to support an intel pc architecture.

This way, Apple reduce the burden on them, their developers and they let the market decide.

 

[firewood]

As long as ARM Macs don't require you to install off the App Store, I'll be delighted.

Even stock iOS devices don’t require you to install apps only off the App Store. You can use Xcode to install any code you can build, just like Linux servers.

 

[skitidetdu]

Windows is always a valid option. I think adobe will port - they are already working on a serious iOS version of many of their apps, after all, and they know they need to clear out the cruft and start with a clean architecture. Some of these apps wont be ported, but the hope is that many more people will buy macs, that developers can easily target the same code base to work well with both macs and the larger installed base of iPads, and so porting becomes more valuable to the developers.
[doublepost=1550818949][/doublepost]

Ok, Americans buy more frequently. There are 200million active iPhones in america. There are 25 million macs. What’s your point?

When did i “focus on mass produced netbooks.” I never mentioned netbooks even once. Why should apple serve their pro users, who don’t make them nearly as much money as iphone users?

Yeah Adobe is getting a head start here. Maybe “valid” was a wrong word to use. If Avid decides to not port protools (just as an example) windows will be the only option. Changing software is much more problematic for workflow than changing what platform it runs on. Maybe this is a none issue in the end...