Apple to End the x86 Mac Era in 2020

[btarunr]

One of the biggest tech stories of the 2000s was Apple's transition from the PowerPC machine architecture to Intel x86, which brought the Mac closer to being the PC it so loathed. The transition wasn't smooth, as besides the operating system, practically every third-party software developer (eg: Adobe), had to rewrite their software for the new architecture, with new APIs, and new runtime environments. Apple could be bringing about a similar change before the turn of the decade.

Apple already builds its own application processors for iOS devices, and some of the newer chips such as the A11 Bionic and A10 Fusion have already reached the performance levels of entry-level x86 desktop processors. It's only a matter of time before Apple can build its own SoCs for Macs (that's not just iMac desktops, but also Mac Pro workstations, MacBook, MacBook Air, and MacBook Pro). That timeline is expected to be around 2020. Since these chips are based on the ARM machine architecture, they will mandate a major transformation of the entire software ecosystem Apple built over the past decade and a half. Intel shares dropped by as much as 9.2 at the first reports of this move.



View at TechPowerUp Main Site

 

[TheGuruStud]

So when this fails are they going to pretend it was the plan all along to turn Airs into useless tablets and switch to AMD?

 

[Ferrum Master]

switch to AMD?

I totally don't understand your point, what are you smoking, where did you get AMD out of this article?

 

[MyTechAddiction]

Well, well looks like intel has some good reasons to crap its pants.
A11 might have entry level x86 performance in the constrained Smartphone format but on a laptop or a desktop with a beefy heatsink + active cooling this might shock everyone.

 

[R0H1T]

So when this fails are they going to pretend it was the plan all along to turn Airs into useless tablets and switch to AMD?

This won't fail, the A11 is already the fastest SoC in it's class & Apple has way more cash then when they moved from IBM to Intel.

 

[TheGuruStud]

This won't fail, the A11 is already the fastest SoC in it's class & Apple has way more cash then when they moved from IBM to Intel.

That doesn't turn it into an out of order performance chip. Also, those benchmarks are fud. It continues to lose in side by side real world tests.

 

[_JP_]

Makes sense, with all the effort they are pulling to keep things in-house and it's an advantage other brands won't reach in the meantime (i.e. Samsung).

 

[FordGT90Concept]

The Intel sell off is a kneejerk reaction because since Jobs died, Apple has been negelecting the Mac line and sales have disappeared. Their market is not growing. It's just the same customers replacing broken Macs with new Macs.

This sounds like MINO: Mac In Name Only. They'll basically be glorified tablets with stands people use to visit the internet. The days of Mac being for the artsy fartsy are over because no one (Adobe, Autodesk, Microsoft, Sony, etc.) is going to rewrite their professional software for a platform that doesn't have the stamina (Apple is liable to about-face again) nor horsepower to be attractive for the customers that made Mac big in the first place.

 

[Fluffmeister]

Yep this makes sense and is clearly the way Apple is going, dumping Imagination Technology of the graphics side, moving to their own powerful ARM based SOCs can remove any reliance of Intel and AMD. Their even developing their own screen technology too apparently: https://www.bloomberg.com/news/arti...o-develop-displays-to-replace-samsung-screens

 

[dorsetknob]

Intel Apple say no
AMD Apple say no
Loongson Apple say no
Arm Apple hint yes

 

[evernessince]

Well, well looks like intel has some good reasons to crap its pants.
A11 might have entry level x86 performance in the constrained Smartphone format but on a laptop or a desktop with a beefy heatsink + active cooling this might shock everyone.

What you're expecting is that ARM will scale perfectly when given thermal headroom, which it simply won't. It would be pretty shocking for ARM to go from bottom of the barrel x86 performance to HEDT performance. In fact it would be one of the biggest leaps in CPU performance ever. You are talking about at least a 400% performance improvement. Given that ARM is pretty mature at this point, those kinds of gains are very very unlikely.

 

[silentbogo]

Apple to End the x86 Mac Era in 2020

LOL. Always love how Apple presents their questionable decisions as "revolution in compute market".
While by the most part Apple SoCs are the best and fastest in the mobile segment, they are still no more than overpowered tablet SoCs.
They are stuck in that one glory-moment when their new mobile chip beat a dual-core i5U on their own macbook pro in a highly-unreliable Geekbench4, and totally forgot that both Intel and AMD have moved on to faster/better things. Even a low-budget 2c/4t Ryzen3 2200u is outperforming that 2017 macbook pro (and that's just the CPU part of perf.), so by 2020 Apple has to come up with something that can not only beat today's low-end x86 laptop SoC, but at least tomorrow's high-end.

The days of Mac being for the artsy fartsy are over because no one (Adobe, Autodesk, Microsoft, Sony, etc.) is going to rewrite their professional software for a platform that doesn't have the stamina

I have a funny prediction: AppStore + iOS ports

 

[evernessince]

Makes sense, with all the effort they are pulling to keep things in-house and it's an advantage other brands won't reach in the meantime (i.e. Samsung).

I don't know if it's really an advantage. Does Apple really have the engineers to sink into R&D for their own processor? Or everything else they want to do in house? I don't really see it turning out well for Apple, you are expecting their in house processors to beat competitors with much more experience and their displays to beat Samsung's. I find it really hard to believe that Apple is just going to magically become the top component maker for processors, screens, and everything else they put in their phones.

 

[silentbogo]

Does Apple really have the engineers to sink into R&D for their own processor?

They do, but they also have a completely different idea of what the laptop CPU should be capable of or be used for.

 

[evernessince]

The Intel sell off is a kneejerk reaction because since Jobs died, Apple has been negelecting the Mac line and sales have disappeared. Their market is not growing. It's just the same customers replacing broken Macs with new Macs.

This sounds like MINO: Mac In Name Only. They'll basically be glorified tablets with stands people use to visit the internet. The days of Mac being for the artsy fartsy are over because no one (Adobe, Autodesk, Microsoft, Sony, etc.) is going to rewrite their professional software for a platform that doesn't have the stamina (Apple is liable to about-face again) nor horsepower to be attractive for the customers that made Mac big in the first place.

Apple has the money to throw at developers, the only problem is if they can make a CPU powerful enough. Right now I don't see that becoming reality by 2020. I doubt ARM will have processors on par with professional level Intel HEDT and by 2020 I wouldn't be surprised to see 10 cores in the x86 mainstream. In addition, ARM doesn't have advanced extensions like x86 does (the disadvantage of being a RISC power saving architecture), which are used by many professional applications.

 

[dorsetknob]

ARM doesn't have advanced extensions like x86 does (the disadvantage of being a RISC power saving architecture), which are used by many professional applications.

Simmiler Argument was made against Apple when they went from Risc (Motorola powerpc ) to X86-64
Apple decided that was the way forward at that time and Software Developers had to write for the new platform or lose out

If they decide Arm is the Way forward then that's that.
as for new Software they either develop for the new O/S (ARM) or they no longer develop for Apple
Apple on ARM is probably where the walled Garden is going ( Apple Probably believe with some justfication that the x86-64 tree is rotten )

 

[efikkan]

There have been rumors about Apple abandoning x86 for many years, and even though we still have no hard evidence, I think they might eventually do it. They are probably at least trying out the feasibility of this transition. Before they moved to x86, they were running parallel builds of OSX for years before they eventually made the switch. When/if they are moving to ARM, they would probably do it the same way.

But keep in mind that Apple might not move their entire lineup to ARM at the same time. The low-end MacBooks have been suggested for years, and will probably be the first to be converted. These machines are too slow to be used for anything productive anyway, and would be a sensible way to test the waters.

I'm more curious what will happen to Apple's more productive computers; the "trashcan" Mac Pro and the new iMac Pro. It's sad to see that the segment which built the company is no longer a priority. If they keep neglecting this segment, it will soon disappear anyway. So they have the options of continuing these on x86, keep them alive with minimal effort until people stop buying them, or discontinue them.

Apple has the money to throw at developers, the only problem is if they can make a CPU powerful enough. Right now I don't see that becoming reality by 2020. I doubt ARM will have processors on par with professional level Intel HEDT and by 2020 I wouldn't be surprised to see 10 cores in the x86 mainstream. In addition, ARM doesn't have advanced extensions like x86 does (the disadvantage of being a RISC power saving architecture), which are used by many professional applications.

The problem isn't mainly that ARM is a RISC architecture, in fact all current x86 implementations are RISC implementations. But as you mention, modern CPUs need various extensions other than pure ALU, FPU and control flow instructions to scale. Some ARM designs does already have some, but are far behind Intel in vector extensions(like AVX, SSE, FMA, etc.) and numerous other instructions which are needed to scale modern applications. The second major factor is the superscalar abilities of a CPU, with multiple execution ports populated by ALUs, FPUs and vector units. The third is the front end of the CPU; the prefetcher with branch prediction etc. A competitive ARM design would need a similar prefetcher, which would increase the die space and energy consumption from current ARM prefetchers. If Apple does all this, then they could be "competitive" with x86 designs to some extent, but not scale beyond it. I still don't consider this very likely by the 2020 timeframe.

 

[MyTechAddiction]

Hedt performance ? No, but maybe something middle of the road.
Considering the apple way they probably already have a working prototype.
They already have the app store and huge number of apps.
Pro users??? well i have no idea.Maybe they will pay adobe to recode their programs.

 

[dorsetknob]

Pro users??? well i have no idea.Maybe they will pay adobe to recode their programs.

Ssorry but thats a laf if Company's Such as Adobe want to sell Software on "The New platform" then they will have ro rewrite/recompile at theit own expense

Hint Did Apple Pay Microsoft to Port Office to the Mac

Spoiler:

OF COURSE NOT Microsoft done it at their Own expense to ensure Sales of Office

 

[FordGT90Concept]

What you're expecting is that ARM will scale perfectly when given thermal headroom, which it simply won't. It would be pretty shocking for ARM to go from bottom of the barrel x86 performance to HEDT performance. In fact it would be one of the biggest leaps in CPU performance ever. You are talking about at least a 400% performance improvement. Given that ARM is pretty mature at this point, those kinds of gains are very very unlikely.

x86 is awesome because it's CISC. Got some new complex thing that lots of people are doing? Add instructions to make it execute in fewer cycles. RISC, RISC basically brute forces everything and at the same time, it's not much of a brute. The only way to get more performance from RISC is more cores and more cores have their own problems (mostly under utilization). ARM is great for mass producing cheap processors that cover the basics. It is not great for much else.

Not to mention, Apple is still going to have to ARM to use it's patents, Apple employees to design and test them, and someone to manufacture them because, AFAIK, Apple is still fabless. What's Apple really saving? On top of that, they're losing a lot of the mindshare that Intel brings to the table because of their constant marketing campaigns. This decision by Apple is full of fail. This is not something Jobs would have ever done. I mean, Jobs maybe would consider AMD because AMD is bringing it now but he wouldn't have authorized changing to ARM for workstations. Workstations are what developers use and without developers, there's no engine driving the Apple wheels other than Apple itself. Jobs would be rolling in his grave if he knew Mac...his original baby...was effectively done for.

I have a funny prediction: AppStore + iOS ports

Exactly what I was getting at, except there won't be any ports. The AppStore will literally be all the new "Mac" can access. Think Microsoft Surface RT. Yeah, that thing that Microsoft discontinued because saving money by using a Tegra processor translated to all that x86 Windows flexibility being lost. Windows Server 2003 for IA-64 too for that matter. Apple got away with changing to x86 because trading IBM for Intel was a huge step up in performance and Intel is synonymous with computing. Apple changing from Intel to Apple will be a huge performance downgrade (hell, probably even slower than the old IBM POWER powered Mac Pros) and Apple isn't exactly known for making good processors.

Apple has the money to throw at developers, the only problem is if they can make a CPU powerful enough.

If "powerful enough" means like 32 ARM cores nothing uses, sure.

 

[windwhirl]

Bound to happen at some point. Though I feel that 2020 may be a bit too soon, they probably will take a while to do the transition, should this be confirmed. Maybe two or three years.

However, Intel shouldn't be too worried yet. This hasn't been confirmed, and even if it were, Apple only brings around 5% of their income. There will be a bit of overreaction at the beginning, but it will be corrected relatively quickly, although Intel's stock value will be a little lower. It's not the end for them, unlike what some people are thinking.

 

[FordGT90Concept]

Hint Did Apple Pay Microsoft to Port Office to the Mac

Spoiler:

OF COURSE NOT Microsoft done it at their Own expense to ensure Sales of Office

Actually Microsoft did it to keep Apple around and keep anti-trust vultures from attacking. Helping Apple in the 90s was mutually beneficial to both companies.

 

[TheinsanegamerN]

This will be interesting to watch.

Everyone claimed apple going for their own CPU design would end in failure. Today, Apple's A series are the best ARM processors bar none, they flat out embarass the best qualcomm can offer.

Then apple went to design their own GPU, everyone claimed it would end in failure. the iphone 8 is supposed to use that chip, and it promises to be quite advanced even compared to previous A series chips.

So if apple says they are going to move from x86 to their own chips, and pulls of the GPU, I'd place bets on them pulling it off. again.

 

[evernessince]

Simmiler Argument was made against Apple when they went from Risc (Motorola powerpc ) to X86-64
Apple decided that was the way forward at that time and Software Developers had to write for the new platform or lose out

If they decide Arm is the Way forward then that's that.
as for new Software they either develop for the new O/S (ARM) or they no longer develop for Apple
Apple on ARM is probably where the walled Garden is going ( Apple Probably believe with some justfication that the x86-64 tree is rotten )

Even if we do assume devs move over it will not happen instantly. How long are creatives willing to wait while devs migrate their code? Given the complexity of modern adobe software, I don't expect it would be quick. Sure you can emulate x86 but the performance is terrible.

 

[R-T-B]

That doesn't turn it into an out of order performance chip.

Last I checked, almost all modern arm designs are out of order, dude.