Apple Plans to Ditch Intel and Use Custom Mac Chips Starting in 2020

[fpnc]

That sort of an iPad Pro is exactly what Apple needs. Chromebooks are eroding the traditional laptop market. Apple can’t afford to let Google have such an unanswered attack on their ecosystem.

You omitted my statement where I said that "I could almost see Apple switching to an A-series processor for something like a Chromebook..."

In terms of whether "we" really need that, I was talking about Mac enthusiasts and those who use Macs as traditional desktops and those who use Macs for scientific, engineering, business, and other pro-like functions. We probably don't need an iPad Pro (with keyboard) or a Chromebook to replace those functions.

 

[Mascots]

A lot of people seem to be forgetting that this means additional money and time from the iOS hardware division - which is literally their bread and butter - will begin flowing into the Mac.

Faster and large upgrade cycles with tighter component integration. It took bringing iOS components to the Mac to get TouchID - this would make that investment more worthwhile for Apple.

They've done phenomenal things over there - let's see some magic for the Mac

 

[rp2011]

Well, so it's beginning. I wonder how many people will choose to move on from Apple due to lack of X86 support. I for one will not be buying a Mac that I cannot run windows on

Windows is as dead as Intel. It just takes a little time for people to realize.

 

[elvisimprsntr]

rereading my post, I think I need to go to bed. I kept swapping PCIE and Thunderbolt...

TI already makes TB silicon. http://www.ti.com/ww/en/analog/tps22985_thunderbolt/index.shtml

Doesn’t seem that much of a stretch for Apple to plunk down a TB chip from another vendor or even drop the TB IP in their own custom silicon.

Apple has a long roadmap and makes small incremental changes which eventually results in game changing advancements. You just have to follow the breadcrumbs. Apple is already conditioning the market with Pro and non-Pro product lines.

 

[lysingur]

It might be possible that macOS in 2020 will ship with some type of emulator (Bootcamp Plus, if you will) if emulation cannot be done at the hardware level. That is, if Apple decides not to have an Intel-based co-CPU on the high-end Macs.

 

[groove-agent]

I thought we were currently in a period without mac updates (Mac Pro, Mac Mini)..

I'm wondering if Apple will be updating their intel hardware anymore. Who will buy a new intel Mac knowing it will be outdated in maybe less then 2 years? Apple used to have rather long transition periods in the past. But how long will they this time support two quite different platforms at massive extra costs. So 3-4 years could be a realistic lifespan for newly bought intel Macs. For companies (depreciation) still "feasible" but for private consumers not very interesting. So will we enter a periode without Mac updates?

 

[tjleonard]

Been covered a lot already.

And no, that’s not what they do

Parallels is virtualization. What would be required is emulation.

Vastly different. Emulation has massive performance trade offs

So everyone is getting bent out of shape, on a rumor, that Apple may actually have a better plan for? Maybe they’re lower cost and they have an intel chipset...maybe a different device all together?

Almost all rumors I read about are not even close.

 

[jinnj]

Its an operating system from a small startup called Microsoft. That OS has a tiny marketshare of 90%, so its quite understandable that you never heard of them

Wait is that the division from Microsoft that was recently shutdown and merged into it's cloud division?

 

[patent10021]

I thought it was obvious when Apple started making their own chips for iOS.

 

[lysingur]

As I said, workload matters. Editing documents all day or writing up code? Not resource intensive. Film editing or photo editing for hours? Gaming? Laptops were never meant to do everything a desktop could do. They were meant to be portable computers capable of doing general tasks while on the go. That combined with the fetish of making everything thinner and lighter results in companies like Apple using minimally invasive cooling techniques so their devices are thinner and lighter, yet suffer.

This shortcoming of laptops has already been solved with native support of eGPU.

 

[groove-agent]

Wha? I don't remember Apple being twice as fast as intel ... certainly during the PowerPC days. Those G4s were so sloooooow.. although the ran nice and cool. 🙂

Can't wait. Apple needs to go back to the days of being twice as fast as PCs, like they were way back when the iMac was introduced 🙂 Intel is such a slowpoke that it shouldn't be hard. About time someone gave those Intel dopes some competition

 

[firewood]

This is what essentially tanked blackberry

“Guys we’re building a whole new os, completely incompatible with what you have today...

Instead, Apple had the foresight to make iOS and macOS nearly identical (except for the mouse-based GUI layered on top in macOS. The Mac desktop is just another app.). The OS kernels, networking, graphics, file systems (except for sandboxing rules), and etc. are nearly identical between iOS and macOS. Either can run on ARM-64 or X86-64. And the mouse-based GUI is just a visual layer than can be easily ported on top of either kernel.

 

[jinnj]

Nope. That lesson was learnt by Steve. Unfortunately he didn't pass that lesson to Tim. So Tim wants to learn it himself again. I would like to see what will be the Enterprise penetration level with the new non-intel Macs.

Enterprise... You missed the memo. But Apple hasn't given a crap about that market in a long long time (Before Jobs returned!).

If it cared, Macs would still have the following ports: ADB, Firewire, SCSI. Hell they would have thrown in a FDC, serial and even a Parallel Port. Thanks Enterprise for moving us forward.

 

[lysingur]

You are comparing mobile to desktop. They are different. very few developers are making any money on the desktop unless the desktop application is for enterprise users. Most home desktop users just use free apps (e-mail, web browser etc.). And since Macs are not used by enterprises (and any businesses for the most part) I am not sure Mac desktop developers will make any money. Add to this the fact that now all the companies selling camera/printer/all devices connecting to computers device will have to develop new drivers for ARM architecture and you get a potential disaster scenario.

Not really. Many cameras and printers already work with iOS through Lightning to USB Camera Adapter, Wifi transfer and AirPrint. Apple is pushing a wireless future so by 2020 there'll be even less need for cables for connecting peripherals, hence completely new drivers for wired devices.

 

[jinnj]

Wha? I don't remember Apple being twice as fast as intel ... certainly during the PowerPC days. Those G4s were so sloooooow.. although the ran nice and cool. 🙂

The G4s and G5s were faster than the Intel CPUs the problem was always cooling them down.

 

[JosephAW]

I actually bought a pre Intel Mac, just to see what it was like (the Mini). But I never considered it for my primary computer until the Intel models came out capable enough for my use. And then it was the 2008 Mac Pro since the iMacs didn't support more than 4GB of RAM.

For my usage (being able to run Windows in a VM, etc), I would think it would be a step backward unless the performance was enough greater that it could run emulation at speed.

The fact it would kill the Hackintosh market, while admittedly probably pretty small, wouldn't be a factor, would it? 🙂

Ummmm my 2006 Mac Pro had 16gb of ram replacing my G5 with 8GB. I think you mean the MacBook Pro had a limit of 4GB( technically6)
Oops. Your right with the iMac. NM 😉
[doublepost=1522721800][/doublepost]

Of course there was the transition from the 6502 to the 65c02 to the 65c816 to the 68000 if you really want to think about transitions. (Although most the 6502->65c816 were pretty easy, from the 65 series to the Mac's 68000 was a big deal.).

My Atari had the 6502 where I learned to code in Assembly. 🙂
[doublepost=1522721853][/doublepost]I love these trial balloons.

 

[rp2011]

There are a lot of folks who own iPhones and iPads who don't own Macs. If this move allows them to use the apps they are familiar with, it could jump-start Mac sales in a HUGE way.

 

[jerwin]

Then, a Mac Pro built on an A-series processor? Not likely.

what about these sorts of chips?

https://www.anandtech.com/show/1202...995-arm-servers-for-cloud-native-applications

 

[firewood]

Being able to efficiently run Windows apps in a VM and being able to run Windows games natively is pretty important to me.

The percentage of high-end gamers who buy new Macs for games is microscopic, too small a number for Apple to care about. Those types buy towers with room for big fans and the latest Nvidia GPUs. Apple caters to iOS gamers, which is already a larger multi-billion dollar software market. They'll grow that market instead, as that's more profitable to Apple than PC games.

Those ancient legacy Wintel apps that high-volume business customers depend on are mostly so ancient that they ran decently on nearly 10 year old cheap Dell laptops, and thus can be easily emulated on a faster new RISC processor.

 

[Freyqq]

If they go this direction, I see the following happening:
1) It will be an ARM-based processor. Macbook will have an ipad ARM processor and Macbook Pro will have an actively cooled 15 watt version of generally the same ARM processor.
2) In the first year, Macbook will be ARM only. Macbook Pro will have an intel processor that it will switch on when running x86 apps kind of like how a GPU turns on when using certain programs. I only see this sticking around for a gen or two.
3) Cost will not decrease, except on the very low-end
4) MacOS will just be the same code-base as iOS except a UI layer with mouse/keyboard support. Apps will run on either one seamlessly except the UI will be changed.
5) Every legacy program will be "broken" and have to be reprogrammed for ARM compatibility. Many programs will never be reprogrammed, mostly games and niche apps. Many of the programs that are reprogrammed will not be as fully featured as the ones lost, possibly for years as they catch up.
6) Apple will lock down the ecosystem just like iOS. All app installs must be through the store, and apple takes their 30% cut. Some applications will never be approved that people may otherwise want to install.
7) iPhones still will not have USB C. iMacs will become docking stations for iPhones to turn into a desktop environment and run the keyboard, mouse, and monitor.

It is an interesting move because it's getting to the point where most of their market is in the iOS realm, so why not just convert it all over to iOS and streamline? It solves a lot of problems. On the other hand, this is a very risky move as they have to rely on developers to redo all their apps or risk a very bumpy migration. They also risk having Intel leave them in the dust as to performance, while Intel is already the standard so no risk in keeping the same chips as competitors. I have a very hard time believing that Apple will come up with a Mac Pro-level Xeon processor replacement using ARM...

Really curious to see how it plays out, but definitely not getting another mac if this turns out to be true. I like dual-booting into Windows and Virtualbox too much to give that up.

 

[lysingur]

Exactly
Power management on ARM has been better with scaling and such.
In terms of performance/watt; Cavium and Qualcomm dissipate about the same type of heat as an equivalent Xeon.
The cost is in the same ballpark also.

People don't seem to get it.
Qualcomm and Cavium are fighting an uphill battle against Intel in the server arena.
They make some in roads in the infrastructure area but they are running Linux and the applications are not off the shelf applications with a GUI.

ARM is not inherently better than Intel in the server arena.
A large amount of your power comes from high speed I/O on the chip.
16 or 32 lanes of PCIe is power hungry.
6-12 channels of DDR4 even if it's LP-DDR4 consumes a lot of power.

An ARM desktop CPU is a lot different than the A11.

Not all desktops are servers. Apple has already moved away from treating desktops as their most important platform and there are very good reasons for that. The future of computing is mobile devices, not only in terms of profitability but also that it is the mobile platform that is situated at the centre of the digital lifestyle, not desktops. ARM chips don't need to be as good as Intel's on the server level to be put into a laptop/desktop.

 

[jinnj]

I find the responses funny. I remember when Apple announced moving to Intel chips and everyone dancing in the aisles. Now Apple is going back to proprietary chips and people are again dancing in the aisles. lol

Except it was not a proprietary CPU, it was a CPU that contained tech (PATENTS!) from Motorola, IBM and Apple. Apple was not able to manufacture CPUs back then so it relied on Motorola and IBM. Motorola didn't give a crap about improving the graphics because it was betting everything it had on communication hardware (now then are part of Google). IBM had no desire to create processors that didn't require active cooling and that affected the Powerbook line (Apple's biggest cash cow at the time). Deal with Intel came in when a Software Engineer got OS X running on Sony PCs starting with the NextStep x86 code. The deal with Intel gave them access to Apple's CPU patents on the PowerPC. Apple may have a similar deal with Intel and if so it will be able to provide On-Chip x86 emulation in it's ARM CPUs. If it doesn't have access to the require patents then AMD could be willing to license the IA-64 instruction set.

 

[PKoz]

While I do have a degree of nostalgia for pre-Intel Macs, I think the incorporation of Intel chips within Apple computers was a net-positive for both consumers and for Apple. Not sure this would be the best move would that it were to happen.

It was absolutely a net positive at the time. It was a much needed transition and they did it. I think it is now a much needed thing to ditch intel. Intel served a purpose. Apple wanted lower power CPU with better performance per watt beyond what PowerPC could deliver. And now Apple wants much the same thing and this time Intel can’t deliver. I marvel at how long my iPad lasts. How long I can leave it in a bag without it getting hot, or using any noticeable amount of power. If it sits in my bag all day and I pull it out, it has essentially the same capacity.

These new Apple chips are amazing. We already have high res displays, high frame rates and excellent graphics. There are more apps on the iPad that I like than the Mac at this point. I’d like to see an Apple CPU powered Mac with a touch keyboard and no need to ever turn it off again. Just let it run and charge it once in a while. That sounds pretty awesome to me.

 

[lysingur]

If they go this direction, I see the following happening:
1) It will be an ARM-based processor. Macbook will have an ipad ARM processor and Macbook Pro will have an actively cooled 15 watt version of generally the same ARM processor.
2) In the first year, Macbook will be ARM only. Macbook Pro will have an intel processor that it will switch on when running x86 apps kind of like how a GPU turns on when using certain programs. I only see this sticking around for a gen or two.
3) Cost will not decrease, except on the very low-end
4) MacOS will just be the same code-base as iOS except a UI layer with mouse/keyboard support. Apps will run on either one seamlessly except the UI will be changed.
5) Every legacy program will be "broken" and have to be reprogrammed for ARM compatibility. Many programs will never be reprogrammed, mostly games and niche apps. Many of the programs that are reprogrammed will not be as fully featured as the ones lost, possibly for years as they catch up.
6) Apple will lock down the ecosystem just like iOS. All app installs must be through the store, and apple takes their 30% cut. Some applications will never be approved that people may otherwise want to install.
7) iPhones still will not have USB C. iMacs will become docking stations for iPhones to turn into a desktop environment and run the keyboard, mouse, and monitor.

3. It's never about cost with the switch. It's about control.
4. This is rather a very good thing, especially for developers. Reasons are obvious.
5. This contradicts with 2. More like there will be a transition period just like you said and might even be possible that Apple will offer emulator to allow legacy programs to be run on macOS of 2020 for the ensuing two years.
6. Apple could've just as easily locked down on the macOS platform now if they want to.
7. By then iPhone will probably have no ports so USB-C or Lightning won't even be an issue.

 

[MacATDBB]

I've worked through Apple's Motorola to PPC transition and its PPC to intel transition and neither of them were good experiences - short term performance hits and critical legacy code inevitably got broken eventually. Who knows what Apple are planning this time, maybe they'll have a unified ISA across devices with some level of short-term HW compatibility for legacy intel x86_64 and ARM ISAs. But this just looks like a world of pain coming back again. Hope this is a trajectory that fizzles for apple R&D.