Apple Killing Intel on Macs? This Is Big

[AndrewFreedman]

Apple is reportedly ditching Intel chips on Macs for its own starting in 2020, which represents a huge shift in strategy and for consumers.

Apple Killing Intel on Macs? This Is Big : Read more

 

[Pnutt79]

For the sake of competition I hope apple does this and it kicks Intel's ass.

 

[dabeargrowls]

Apple is slowly destroying itself. They dumped the powerpc and pissed off a lot of people (me included) for Intel and now they are going to do it all over again? haha...dont fix what ain't broken! Intel and AMD are experts and these apple yahoo's can't even build a phone worth a damn. Wireless charging 5 years after its been out... that's a breakthrough! I'm all for it though because I want to see them destroy themselves since Steve left. No vision whatsoever.

 

[weilin]

I'll copy over what I wrote on Engadget...

I have my doubts on this 'rumor'... This would be an expensive transition, and I'm not sure apple stands to gain enough to do this (certainly not financially). Though, if anyone were to try (and succeed), it would be Apple. Of all the ARM companies around, they are the most equipped to estimate the trade offs to pursue this.

It's not that ARM ISA can't scale to desktop, it's just no one has bothered to do it yet. One of the biggest reasons why is you would need to sell a stupid number of chips a year to make this profitable. The closest I can think of is AMD, a fabless CPU company (lets ignore the GPU part). Even with their volume, they were still fiscally negative for years! This isn't something most companies can blindly invest in. It would require a team of thousands to make a desktop class processor. The time to market is pretty long so they would need to have started in 2017 (honestly, it should be pretty close to completion by now) to have something ready by 2020. Then they have to keep that pipeline fed to have yearly refreshes.

Emulation can be done (see Windows on ARM with x86 emulation), However, initial reviews shows it comes with an insane performance impact. At this point, you're better off with an iPad Pro running native code than emulation.

Binary backwards compatibility would initially be an issue, so I looked up the top apps on OSX (and the top professional apps on OSX), and most (all?) companies look like they're doing well so at least there's someone around to port the most popular ones to the new ISA. Most of the companies are small (few products) and a few are corporate giants. For the former, they must adapt or die, for the latter, they'll just do it because they have to. Given how (relatively) new the codebase is for x86 OSX, I'm willing to bet most of the source code is still around and (given good software practices... ahem...) easily portable.

Not to mention, if apple does this, all first party apps/suites/etc. would be ported from the get go. If your app is important enough, it wouldn't surprise me if you got an emulator/dev unit to port your applications before the official release date. If you're a big company like Adobe, I'm almost certain you will something before everyone else.

My bet is, if this actually happens, average user probably won't even notice that the ISA changed. I think this ISA transition will go smoother than the PowerPC to x86 one.

Hackintoshs/BootCamp/Etc. would all be in doubt...

 

[mrmez]

Apple is slowly destroying itself. They dumped the powerpc and pissed off a lot of people...

Totally agree. Just look how badly Apple has struggled since it dropped PPC in 2006.

I'm still devastated we don't have 5.25" floppy disks anymore :`(
I mean... why do we need to change anything? Ever???

 

[shrapnel_indie]

Apple started with CISC (Apple I & II series 6502. GS customized 8/16 bit 6502 type, Liza/Mac - Motorola 68k series) to RISC (IBM PowerPC) back to CISC (Intel x86/x64) and now talking back to RISC? All this flip-flopping around....

Yeah, their finances (and stockholders) required changes to take place AND those changes hurt some further than the hurt they were already in... but seriously.

 

[therealduckofdeath]

This would probably spell the end for their PC platform. Then again, it does seem like they don't really care about it particularly much. With the lack of support for all contemporary graphics API's and such. ARM is nice, but it's not competitive with x86 on performance. Not the same ball park, not even the same sport. Just look at how terrible PUBG performs on mobile, despite being low detail, upscaled graphics. On mobile it works for Apple with a slightly differentiated ARM design, as the power draw drastically limits design and innovation. Not on a platform where power draw and heat generation is irrelevant.

 

[Dosflores]

This isn't going to be an expensive transition. If they do the transition, it's because they expect to save money in the long term. It's as simple as that. No one would stop using a powerful processor and start using a weaker one for any other reason than to save money.

It's not going to be difficult for Apple to perform, since they have already made it once. They have already developed all the tools needed for a CPU architecture switch.

They don't care about backwards compatibility. They can easily tell every Mac App Store developer to update their apps to work on ARM. If they don't comply, they'll be thrown out of the Mac App Store. It's not going to hurt Apple.

And what about the apps you get outside of the Mac App Store? Apple cares even less about them. They don't make any profit out of them, anyway. Sure, they're appreciated by professional users, but how many MacBook buyers are really professional users? They can put the Pro branding on anything they like, but I don't think any professional that needs top performance would look at a MacBook Pro and think "that's a great professional laptop".

So, Apple is going to keep making lots of money, as usual. The opinion of people who bother to write on technology forums is totally useless for them. They're not their target customers.

 

[manleysteele]

Where do these analysts get this meme that Intel's business is on the server side? Client computing had more sales than all other sectors combined again last year as they have for many years. That includes server SKU's which was the second largest sector, but trailed client computing by a significant amount. Don't these "analysts" know that we have access to the same numbers they use? If their was a huge difference in profitability of the two sectors that would explain it, but the difference in margins is small.

 

[redgarl]

Why would they do that? They still need GPUs for their platforms. It is going to be quite risky for nothing.

 

[hannibal]

Maybe air models go to arm.... cheap 1000$ devices...
And pro models could use Intel at least at this moment From 2000$ and above... it could be usefull to product segmentation purposes. But only if They use same cpu in iPads, phones and air laptops.
As someone did say, it is not cheap to make new cpu so I am very sure that mobile devices and new laptops share the same cpu. And mobile cpu is fast enough to normal Office tools and websurfing. More demanding can be done with pro series devices.

 

[therealduckofdeath]

They don't care about backwards compatibility. They can easily tell every Mac App Store developer to update their apps to work on ARM. If they don't comply, they'll be thrown out of the Mac App Store. It's not going to hurt Apple.

This is not mobile. Apple has a diminishing market share on desktop, way down in the single digits today. Developers don't just churn out completely re-written applications just because an operating system developer tells them to. If this is the thought process of Apple, they're going to get a cold awakening when they realise they'll barely get any third party desktop applications.

 

[Dosflores]

This is not mobile. Apple has a diminishing market share on desktop, way down in the single digits today. Developers don't just churn out completely re-written applications just because an operating system developer tells them to. If this is the thought process of Apple, they're going to get a cold awakening when they realise they'll barely get any third party desktop applications.

There's no need to completely re-write applications. The macOS ARM API should be 99% identical to the x86 API. Developers would only need to recompile and test a little. Not a big deal, and totally worth it in order to get a few more app sales from new MacBook buyers.

 

[toadhammer]

I think they still have a large share of sound and video markets. But they have long treated desktop machines as an afterthought, and it's even less surprising that the desktop impact is very unclear in this rumor.

Still, I believe this is all about laptops, or at least the non-pro laptops. With Surface as a model, there's not that much difference between the light laptops and an ipad pro. And if you follow Dvorak, this is something he's been predicting for years.

 

[toadhammer]

I'll also point out there's a world of difference between desktop-only applications and "apps" that run on laptops or ipads. It's really not feasible to switch some of those applications over to ios, or at least not without major hoopla. I think the staggered rollout applies only to laptops running ios. Any plan they have for desktops would have to lag that, or involve dropping them outright, or keeping them as-is.

 

[Dosflores]

I'll also point out there's a world of difference between desktop-only applications and "apps" that run on laptops or ipads. It's really not feasible to switch some of those applications over to ios, or at least not without major hoopla. I think the staggered rollout applies only to laptops running ios. Any plan they have for desktops would have to lag that, or involve dropping them outright, or keeping them as-is.

The rumor doesn't mean that iOS will be ported to Macs, but rather that macOS will receive an ARM version. It's what makes the most sense, since Microsoft has already released a version of Windows 10 for Snapdragon processors.

iOS will remain exclusive to iPhone and iPad, and macOS will remain exclusive to Macs. However, both operating systems share their kernel (just like GNU/Linux and Android), so Apple can add some compatibility layer in order to be able to run iOS apps on macOS.

 

[oneblackened]

Yeah, I'd be less than surprised if it takes them longer.

ARM is optimised for low power, and it doesn't scale up to non-ULP x86 levels of performance particularly well, plus god knows Apple isn't gonna invent an ISA.

 

[alextheblue]

Yeah, I'd be less than surprised if it takes them longer.

ARM is optimised for low power, and it doesn't scale up to non-ULP x86 levels of performance particularly well, plus god knows Apple isn't gonna invent an ISA.

That's not because of the ISA. There's no technical reason you couldn't build a higher-power, higher-performance ARM chip. The main reason to avoid doing so is that it would put you in direct conflict with AMD and Intel, who both have a lot of experience building high performance chips in the ~10-15W and up range. It's a massive undertaking that would mean a new architecture designed for higher power. But if you're Apple and you want to round out your vertical consolidation efforts, it makes sense. Plus they already have decent-performing SoCs in their iPad Pro. Starting with the lowest-power MacBook Air models, it wouldn't be that hard for them to start gradually building higher power SoCs for their Mac lineup.

Plus they've changed architectures successfully twice that I can recall. From Motorola to PPC and then to Intel. They're pretty good at it. Between emulation, native compiled libraries/APIs, and recompiled software they'll have their bases pretty well covered. Oh, and before someone calls me a fanboy I have never personally owned an Apple product. I don't care for them as a company but they do build some decent hardware from time to time, even if it is overpriced.

 

[bit_user]

Where do these analysts get this meme that Intel's business is on the server side?

Margins are much bigger in server/datacenter and cloud is a growing market. Meanwhile, the PC (i.e. desktops & laptops) is shrinking with no end in sight.

If their was a huge difference in profitability of the two sectors that would explain it, but the difference in margins is small.

According to whom?

 

[manleysteele]

Where do these analysts get this meme that Intel's business is on the server side?

Margins are much bigger in server/datacenter and cloud is a growing market. Meanwhile, the PC (i.e. desktops & laptops) is shrinking with no end in sight.

If their was a huge difference in profitability of the two sectors that would explain it, but the difference in margins is small.

According to whom?

According to me. I read the numbers directly. You can, too.

 

[richardstaller]

Apple moving towards the ultimate and most profitable consumer platform while leaving the pros and creatives who use their computers behind.

 

[cryoburner]

It is no secret that iOS has significantly larger volume of apps available than macOS, so if Apple can get more of the iOS developers to create apps for macOS without much additional work, it could be huge for the Mac app ecosystem and the Mac platform in general.

I'm sure Mac Pro users will be thrilled at the wide array of Flappy Bird clones and finger-painting apps this will open up to them. : 3

 

[bit_user]

It is no secret that iOS has significantly larger volume of apps available than macOS, so if Apple can get more of the iOS developers to create apps for macOS without much additional work, it could be huge for the Mac app ecosystem and the Mac platform in general.

I'm sure Mac Pro users will be thrilled at the wide array of Flappy Bird clones and finger-painting apps this will open up to them. : 3

I wouldn't be so dismissive. Once you can run iOS apps on MacOS, app developers will probably add depth and breadth to many apps.

MS tried to move towards a unified platform with Windows 8. Google is doing it with ChromeOS. I don't think it's such a crazy idea.

 

[cryoburner]

I wouldn't be so dismissive. Once you can run iOS apps on MacOS, app developers will probably add depth and breadth to many apps.

MS tried to move towards a unified platform with Windows 8. Google is doing it with ChromeOS. I don't think it's such a crazy idea.

What works well on mobile does not necessarily translate well to the desktop though, and the opposite is also true. Windows 8 should be a prime example of that. Or look at a majority of games that get developed for both desktop and mobile. More often than not, the kinds of interfaces that work best for smartphones and tablets don't translate particularly well to a mouse and keyboard, and vice versa. And that even extends to features, and what people are looking at using a program for. Smartphone apps are typically used for short bursts on the go, while desktop software is more likely to be worked with for extended periods, and that can affect what kind of features make sense on a given platform.

 

[bit_user]

What works well on mobile does not necessarily translate well to the desktop though, and the opposite is also true.

*sigh* nobody is saying desktop-only apps will go away. You're using an extreme position to argue that a firewall be maintained between mobile and desktop/laptop? Nice.

Windows 8 should be a prime example of that.

Well, you know, Apple's Newton was a failure. So, I guess people really don't want handheld computers.

Sheesh. Just because someone tries something and it doesn't turn out great doesn't mean the idea has no merit or that a better execution isn't possible. This is MS, we're talking about.