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

[usarioclave]

The question is, how serious is Apple about its chipmaking? Can they do that for 15 years?

Just look at SPARC and how it basically died on the vine. Chips need a ton of R&D and continuous improvement. The A-series has shown that Apple can do it, but do they have the willpower and budget to make a desktop-class processor? Is the A (or B-series) good enough?

There are plenty of sneaky things they can do to make the transition, like have multiple processors and a lot of pre-building behind the scenes. But putting Apple chips in basically means the death of Apple's prosumer product as we know it.

 

[RaoulDuke42]

Nice to get rid of x86. I won't buy it. Apple should have not abandoned PPC.

Soooo... you’re still rocking a G5 or some such? ‘ardcore!

 

[MichaelDT]

People don’t buy macs so they can run windows. Almost nobody does it.

Except every developer.

Edit: I depend on my ability to boot and/or virtualize x86 versions of Linux and Windows. Servers run x86 Linux, buisnesses mostly run Windows for desktops. How are software shops going to test what we develop on Macs? TBH most will just switch to Wintel machines and boot Linux as the primary.

 

[cube]

Soooo... you’re still rocking a G5 or some such? ‘ardcore!

Apple forced me to the dark side. Twice. Now I use a "gaming" laptop.

Intel because of Thunderbolt. Make that three times. OK, Intel also because mobile CPUs are more powerful than AMD's still (if you get discrete graphics).

And it has NVIDIA too (not by choice).

 

[Lozza013]

According to rumors, transition to ARM will be over several years, starting with low power devices, giving diehards opportunity to get an Intel machine that will be powerful enough for foreseeable future. Mobile and low power devices is the future where ARM will shine, including car electronics and home automation

If that turns out to be the case, I'd be more than happy with that.

 

[ArrayDecay]

I don't believe Apple sees Macintosh laptops, desktops, and power machines in the long-term future.

 

[Michael Scrip]

I think by Apple making their own SOCs, they can be cheaper than by buying premium chips from Intel, and AMD would probably be able to do custom for fairly cheap—they make console APUs for cheap.

As others have said... Apple will have to pay off all that R&D for creating this completely new generation of processors.

Sure... Apple won't be buying $400 processors from Intel anymore... but the cost isn't suddenly zero.

And I also don't think the price of the Mac will suddenly drop because they're using their own homegrown chips.

If anything... Apple will convince people to pay more for Macs since their new processors have an even better connection with MacOS. Integration, ecosystem, and all that jazz.

 

[cmaier]

Except every developer.

Edit: I depend on my ability to boot and/or virtualize x86 versions of Linux and Windows. Servers run x86 Linux, buisnesses mostly run Windows for desktops. How are software shops going to test what we develop on Macs? TBH most will just switch to Wintel machines and boot Linux as the primary.

Not every developer.

Again, some people need it. Just not enough to matter. Even if every single person who wants to dual boot never buys another mac, the number of NEW mac customers because of this change will outnumber the sales lost by two orders of magnitude.

 

[abrantes09]

Are those i7 laptops you're using in this statement?

No, an A11 iPhone. Nice try though

 

[ASentientBot]

If this happens, it's likely the end for Hackintosh and our unsupported Macs... As well as hundreds of apps that'll never get updated by the developers. It'll be fine in the long run, just like switching to USB-C, removing headphone jacks, etc. probably will be too. But just like those things, it's going to be a real hassle in the meantime and they'll lose a lot of users.

 

[derek]

Nope! Not gonna happen. Here are a few reasons why:

1) No more virtualization of Windows on Mac.
2) No more directly running Windows on Mac (aka Boot Camp).
3) No way Intel is going to license their CISC proprietary technology to Apple to run on A-Series RISC CPUs.
4) Most Mac software would be dead-ended and would have to be substantially rewritten to run on RISC CPUs.
5) No more macOS = UNIX without a major rewrite of UNIX for RISC CPUs.
6) There is at this time no realistic speed advantage for moving from Intel CISC chips to Apple A-Series RISC chips.
7) No more GPUs for Mac without again a substantial driver rewrite, if that's even possible with the loss of proprietary Intel CISC technology. Apple has screwed things up badly for Mac GPUs already, locking out anything new by Nvidia, aka stupid move.

And there's more. But I find the above is enough to freak out most Mac users.

And please, those who don't understand the difference between CISC based Intel CPUs and A-Series RISC based CPUs, please read about the difference before making statements that prove your ignorance. Thank you. Oh and no, Apple's Marzipan project is entirely unrelated to this stupid rumor.

The real Question: Why does this stupid rumor come up year after year? IMHO it's re-perpetrated by those who have no idea of the technology shift involved or the damage that would be done to the Mac market if it happened.

 

[Woochoo]

Many people here just seem to unknowledge that most apps are easily compiled (not ported) to ARM from the IDE itself, and it's basically the same app without running over any fancy layer so no overheat. But also Apple is planning to move to universal binaries, so what's the deal with all those people crying about running on emulators?

The second thing people just seem to unknowledge is that actual ARMs are crushing many dual cores (i5/i7) used on ultrabooks both in performance and perf/watt, and I'm not even talking about Apple's custom designs (that usually are leading). And don't come with the "buh...buh u can't compere ARM with x86 benchmarks!". In the end both ARM and x86 have a RISC pipeline, the only big difference is x86 has a decoder for complex instructions but it's not even pure CISC, but they end up executing stuff in a similar way.

 

[Suttree]

I gave up on Apple when they switched away from PowerPC. Third party apps I used never got their Intel counterpart.

 

[Darmok N Jalad]

I don’t think they will lower prices for consumers after building their own SoC.

I don’t mean to imply they will make a product cheaper. Instead, more margin goes to Apple than to Intel.

As others have said... Apple will have to pay off all that R&D for creating this completely new generation of processors.

Sure... Apple won't be buying $400 processors from Intel anymore... but the cost isn't suddenly zero.

And I also don't think the price of the Mac will suddenly drop because they're using their own homegrown chips.

If anything... Apple will convince people to pay more for Macs since their new processors have an even better connection with MacOS. Integration, ecosystem, and all that jazz.

Oh no doubt, but look at Apple’s cheapest non-iPhone. It’s a $329 iPad running older Apple SOCs. They can’t make a Macmini for under $799. Apple will have to invest into their own chips, but they don’t have to pay Intels fee for the same thing

 

[nouveau_redneck]

If a Mac Pro is coming around September, and a few months later they're switching everyone to ARM, that's bad timing. Or the Mac Pro may be the first computer with an array of ARM processors.

Perhaps the Mac Pro modular form will accept both Intel and ARM modules via whatever mechanism makes it modular. So in that respect if the new Mac Pro was released in 2019 with Intel, the same chassis could be upgraded when the time came that Apple has ARM ready, and the customer was ready to use it.

When that point comes, perhaps they offer either Intel or ARM as a choice for a period of a couple years, so the customer could chose based on needs during the transition years.

 

[cmaier]

Nope! Not gonna happen. Here are a few reasons why:

1) No more virtualization of Windows on Mac.
2) No more directly running Windows on Mac (aka Boot Camp).
3) No way Intel is going to license their CISC proprietary technology to Apple to run on A-Series RISC CPUs.
4) Most Mac software would be dead-ended and would have to be substantially rewritten to run on RISC CPUs.
5) No more macOS = UNIX without a major rewrite of UNIX for RISC CPUs.
6) There is at this time no realistic speed advantage for moving from Intel CISC chips to Apple A-Series RISC chips.
7) No more GPUs for Mac without again a substantial driver rewrite, if that's even possible with the loss of proprietary Intel CISC technology. Apple has screwed things up badly for Mac GPUs already, locking out anything new by Nvidia, aka stupid move.

And there's more. But I find the above is enough to freak out most Mac users.

And please, those who don't understand the difference between CISC based Intel CPUs and A-Series RISC based CPUs, please read about the difference before making statements that prove your ignorance. Thank you. Oh and no, Apple's Marzipan project is entirely unrelated to this stupid rumor.

The real Question: Why does this stupid rumor come up year after year? IMHO it's re-perpetrated by those who have no idea of the technology shift involved or the damage that would be done to the Mac market if it happened.

Apple doesn't care about 1 or 2. 3 is irrelevant - apple doesn't need an intel license. 4 has happened multiple times before (68k to ppc, ppc to intel, and 32 bit is about to go away) and we survived. 5 is nonsense. iOS already runs on a similar kernel to macOS, and macOS already undoubtedly has been running on ARM for years internally to Apple. 6 is wrong. There is no technical reason ARM has to be slower than x86, and by getting rid of the large instruction decoders and complicated addressing modes, more transistors can be dedicated to payload, clock can be turned up within the same thermal envelope, or both. 7 is a potential issue, but they've already been pushing everyone toward Metal for that reason. CISC has nothing to do with graphics.

 

[Reindeer_Games]

Not every developer.

Again, some people need it. Just not enough to matter. Even if every single person who wants to dual boot never buys another mac, the number of NEW mac customers because of this change will outnumber the sales lost by two orders of magnitude.

Care to share your source for that market research? You have an educated opinion, I'll admit-but unless you are Apple, your numbers are just subjective, educated, opinion.

 

[cube]

Not every developer.

Again, some people need it. Just not enough to matter. Even if every single person who wants to dual boot never buys another mac, the number of NEW mac customers because of this change will outnumber the sales lost by two orders of magnitude.

They could just sell sugared water if all that matters is volume.

 

[Michael Scrip]

Oh no doubt, but look at Apple’s cheapest non-iPhone. It’s a $329 iPad running older Apple SOCs. They can’t make a Macmini for under $799. Apple will have to invest into their own chips, but they don’t have to pay Intels fee for the same thing

Yeah... Apple can make their own processors for half of what Intel charges them...

AFTER they spend a billion dollars on the whole thing.

But you're exactly right. We already know how powerful iPhones and iPads are with their own chips. Maybe their laptops and desktops will follow the same trajectory.

I'm interested to see what they do with this. And how they handle any software compatibility/performance issues.

 

[cmaier]

Care to share your source for that market research? You have an educated opinion, I'll admit-but unless you are Apple, your numbers are just subjective, educated, opinion.

That’s all they are. Educated opinion based on the fact that apple has a huge base of potential customers in the people who own iPhones and iPads and want an easy way to run that software on the desktop, and the quantity of those people is far larger than the entire install base for macs and therefore larger than the number of people who use macs for windows.

 

[Anarchy99]

Personally I doubt this is true, I think it's more likely we have a few years of Mac having a powerful ARM co-processors.
then eventually a move to ARM years later maybe if that goes well.

It would really come down to how well they can emulate X86.
Which has own issues and unlike with rosetta i don't think Transitive technologies exists anymore they did have a binary translation solution for x86 to ARM but i seem to remember intel getting into a legal dispute over patents involving x86.

then even in that hypothetical scenario even if you Apple such a translation layer that handles old mac Apps but it's common for people to use windows on a mac either native via bootcamp or for the less technical users, parallels or vmware etc.

so would we then be talking about apple needing a "Classic (x86) Environment" because the challenges for a full emulator are much more involved than binary translations.

the way i see it is even worse case scenario if every mac got changed to arm in 2020 without exception they'd still support intel macs for 4-8 years or so
so it will be back to hackintoshing for me.
then when that runs its course in a few more years i could re-evaluate.

 

[cmaier]

They could just sell sugared water if all that matters is volume.

The good news is that by switching to custom ARM chips they not only increase volume but make the machines better.

 

[Reindeer_Games]

That’s all they are. Educated opinion based on the fact that apple has a huge base of potential customers in the people who own iPhones and iPads and want an easy way to run that software on the desktop, and the quantity of those people is far larger than the entire install base for macs and therefore larger than the number of people who use macs for windows.

So you say you have a Doctorate in Applied Science but you can't see the leap of faith you are taking in the presumption of your hypothesis that a pool of untapped marketshare of iOS users would spend more money on a product that was only marginally better?

My response-let the experiment commence.

 

[Wilderness-1902]

think you're looking at that the wrong way around: currently a developer who wants to support both iOS and MacOS has to write and maintain two different apps - many do, but they must be continually looking at the relative market sizes and wondering if the Mac version is worth the effort. Being able to target both with a single app would make them more likely to support (or continue supporting) the Mac.

This will make iOS development more difficult, since we need to support additional hardware configurations.

I don't think moving to ARM will boost Mac's sale. If developers cannot make enough money, the compatibility of universal apps will become a burden instead of a good feature.

A similar case is there are a lot of work to make a web app run well on both desktop and mobile Safari

 

[cmaier]

So you say you have a Doctorate in Applied Science but you can't see the leap of faith you are taking in the presumption of your hypothesis that a pool of untapped marketshare of iOS users would spend more money on a product that was only marginally better?

My response-let the experiment commence.

They already do that when they buy iPhones. They ain’t cheap.

And it’s a Phd, a JD, a BS and a MS.