Mac Minis with ARM Prototype chips.

[keysofanxiety]

Apple is wasting a lot of the potential of OS X by limiting it to so few devices. I think OS X should be standard operating system people use on practically any device; it really is that good. But it sure looks to me as though Windows (or Linux) users are going to continue to increase at the expense of OS X users, as there are just so many more platforms available for those operating systems.

I really disagree with this. Apple's strength comes from optimising/designing for specific hardware. They did consider this back in the early 2000s with OS X running on Sony Vaios (we know how much Steve loved them), but it never materialised. I bet Sony are kicking themselves about that!

Right now the rMBPs are the best power portables on the market in terms of build quality, thinness, lightness, battery life... Apple have almost complete creative control over everything they do. The trackpads are unrivalled. They'd just look to tighten that in future to ensure quality and standardisation, not loosen it. Complete control over the CPU pipeline and in-house development for OS X would certainly be a consideration.

One of the many, many problems with outsourcing OS X would just eventually mean quality would suffer, as OEMs do a race to the bottom.

If there was ever a valid time for the argument of OS X on non-Apple systems, it would have been back in the mid-2000s when their marketshare and sales were much lower. Not now. Definitely not now.

 

[Omega Mac]

Interesting debate. I think the main thing now is to watch for more signs and signals of Apples real longterm roadmap that points it toward ARM.

If they are pitching their A9X iPad as a super computer, I think you can bet they are moving to put them in their computers. They are opening up the market to the idea of A-Series powered Mac's. Apple are doing a lot of talking about their Chips. People are using them in their iPhones and iPads and the iPhone6 form my experience is amazingly capable device.

I do not believe any other phone maker is in such a position to remodel the market demands in their image. This reminds me of what Steve Jobs didn't see in the Amiga (as he was given an early look but passed) but has arrived there anyway, they are going right back to that whole package approach.

I also cite the recent Intel / Arm foundry deal. Intel are Pivoting, I think they know where the market is going and a big customer is going to want more the iPhone ARM silicon (maybe of a more demanding variety as required in traditional computers) so maybe they'd like a slice of the iDevices too in the long run as par tot a partnership.

Surely Intel have the best fab expertise in the world? The kind you need when producing stealer class A-Series Chips

 

[Crosscreek]

If Apple moves to ARM they will in essence lock out all professional work as far as X86 applications requiring high video compute is concerned. While Apples ARM chips has a great video controller on board it will never match discreet video card compute and power.

 

[hawkeye_a]

I have to admit that "prototype chips" could also mean new intel chips. They could just be testing how they cope and handle inside the mini form factor. (Would be pretty standard testing for unreleased products)

i doubt the new CEO would be able to orchestrate such a transition anyway, and would probably avoid.

From a business standpoint, there is no real reason at this point for Apple to invest in such a transition. They have a stable partner for chips and the Mac is a much smaller part of their business today when compared to 2006.

 

[Omega Mac]

I have to admit that "prototype chips" could also mean new intel chips. They could just be testing how they cope and handle inside the mini form factor. (Would be pretty standard testing for unreleased products)

It' simple to deduct, ARM + iDevices + Genius Silicon Wizard

The entire article focuses on the man behind the A-Series Chip and the A-Series in terms of it's monumental transformation of Apple's fortune.

For the majority of people on the planet Apple is iPod but now more so iPhone.

In pops some mini's making a cameo bearing prototype chips, now what kind of prototype Chips is the A-Series maker going to be making in his never before seen lab?... yes that's right next gen A-Series next chips.

Focus people, FOCUS!

 

[jpietrzak8]

I really disagree with this. Apple's strength comes from optimising/designing for specific hardware.

Well, fine. They can go ahead and optimize / design a desktop computer that works great for anyone doing light office work or web browsing.

Since I'll be using Linux on my dev boxes (a computing niche that Apple ignores), and using Linux on my gaming box (another computing niche that Apple ignores), if I really need or like the Apple hardware I'll go ahead and buy one and use Linux on it. There's no point in me having to support multiple operating systems when just one will do the job fine.[/QUOTE]

 

[keysofanxiety]

Well, fine. They can go ahead and optimize / design a desktop computer that works great for anyone doing light office work or web browsing.

Since I'll be using Linux on my dev boxes (a computing niche that Apple ignores), and using Linux on my gaming box (another computing niche that Apple ignores), if I really need or like the Apple hardware I'll go ahead and buy one and use Linux on it. There's no point in me having to support multiple operating systems when just one will do the job fine.

You should absolutely use the OS and computer that you feel is best! Personally I feel that OS X wouldn't really be for me unless it was on the hardware it's on. I like the OS but I wouldn't instinctively snap to use it on a different machine. So it's more of a mix of hardware & software, which is partly why Windows/Linux doesn't appeal as much as it used to.

I think it's a little unfair you began by judging Macs as just for web browsing. The high-end systems are still extremely capable computers, even if they are priced a touch heavy on the wallet, and Apple honestly do some great bespoke creative software.

 

[Bubba Satori]

What computer operating systems and software will run on these ARM computers?

 

[rGiskard]

ARM Macs would create more problems than they would solve. Namely, Macs would lose functionality, since they could no longer run Windows natively. This would be a dealbreaker for many professionals who require a few Windows-only apps.

Then there is Intel. It is one thing to take on Samsung and Qualcomm, but an entirely different proposition to take on Intel. ARM solves a problem: how does Apple make fast custom SOCs for it's mobile devices. What problem would an ARM Mac solve? Intel already offers the fastest desktop and laptop silicon in the world.

Macs are a very small portion of Apple's business. Why would they invest extra Billions in ARM to compete with Intel in a shrinking market?

Maybe Apple could develop some sort of complex Mac with both Intel and ARM silicon that could run OS X, Windows, and iOS, but again, what problem would such a machine solve?

 

[Algus]

Sharing chips with the Windows world has done us a lot of favors in terms of the types of software available on OS X. Software availability has always been an issue on Mac OS vs Windows because of the smaller market. I am not particularly eager to move to a different processor family again regardless of whatever performance incentives such a move might offer.

For some years I ran an ARM Chromebook and even this caused issues as while Google did a good job of getting ARM to run everything an Intel Chromebook could, when the time came for me to run a Linux distro, there was plenty of software unavailable to me that ran just fine on an Intel build of the same distro.

I am sure they have been experimenting with ARM and are looking ways to cross platform iOS and Mac OS. I have not been particularly pleased by the obvious influence iOS has had over Mac OS and for me a move to ARM and Macs getting the ability to run iOS apps would not be a satisfactory alternative to Intel compatibility. Ultimately it would move me off of Apple computers and back to Windows machines.

I strongly suspect that if Apple decides to move to ARM they won't care about chasing users like me anymore.

 

[BeefCake 15]

ARM Macs would create more problems than they would solve. Namely, Macs would lose functionality, since they could no longer run Windows natively.

Windows 10 has started a light version to run on ARM which is a huge move from Microsoft to support this platform. They're looking at ARM and possibilities with it especially since they want one OS to implement everywhere.

 

[hawkeye_a]

@Algus the availability of software would definitely be an issue. Apple could optimize all its software to run off ARM but it would take time(and money) for others. Something they could decide not to do or support cause of the tiny marketshare.

Moving from 68k to PowerPC or PowerPC to intel, there was a significant "step up" which made emulation an option. However moving from Intel to ARM, at least now, wouldnt be kind to emulation.

 

[rGiskard]

Windows 10 has started a light version to run on ARM which is a huge move from Microsoft to support this platform. They're looking at ARM and possibilities with it especially since they want one OS to implement everywhere.

People who buy laptops will often need the real version of Windows, not some gimped joke. When Apple's ARM implementation features more processing power than Intel's desktop/mobile silicon then sure, maybe it could be worth it. Until then, no way does Apple invest so much in a dwindling market. They won't even put Intel's latest silicon in their Macs, so why would they spend billions to develop their own silicon?

 

[grahamperrin]

… may be way off, is it possible that the new file system they're developing APFS a precursor to the introduction of ARM on a Mac? …

Probably way off (I don't see the storage system, and ARM, as in any way linked on a timeline).

Discussion: Introducing Apple File System (APFS)

 

[fischersd]

I hate to be the one to tank the discussion, but the mini's could just be running the automated testing, connected to the sysbrd's that have the ARM processors on them. Why mini's? Gee, we have thousands of them in the warehouse that aren't selling. . It's a pretty low cost testing platform for them. Depending on the nature of the testing, there may not be economies of scale you'd get from a decent sized VM computing farm.
Sorry.

 

[Flynnstone]

There is no question, in my mind, that Apple is running ARM based Macs in the lab. They've done it before. It's in front of us. iOS can be considered Mac on ARM.
x86 is banging on limits. Microsoft has Windows on ARM.
One way to get a feel for this situation is to measure Intels pulse. Are they freaking out.
Well, there are talks of Apple making deals with Intel to fab ARM devices for Apple.

 

[jpietrzak8]

x86 is banging on limits. Microsoft has Windows on ARM.
One way to get a feel for this situation is to measure Intels pulse. Are they freaking out.
Well, there are talks of Apple making deals with Intel to fab ARM devices for Apple.

This reminds me, way way back, RISC processors were just getting hot, and everybody was saying the same thing: x86 architecture is fundamentally flawed! Intel cannot compete! The new RISC architectures are the wave of the future!

That was, what, 20 years ago? Seems the future has yet to arrive.

 

[danmart]

Just to give another piece of the ARM picture, running industry-standard servers on ARM is not a new thing. HP have been doing this for a few years now (link). I had the benefit of going to an HP industry event a few years back when this stuff was still in development and the chips were faster, cheaper and more energy efficient than the Intel equivalents, so a fairly compelling proposition!

This doesn't mean that Mac will be going ARM any time soon, of course, and I agree that moving to Intel chips was a big plus for Mac in terms of easier porting of apps from Windows to OSX, so being first-movers here might be unwise.

 

[theluggage]

This reminds me, way way back, RISC processors were just getting hot,

(Actually, no, it was Intel processors that were getting hot: RISC processors were cool!)

and everybody was saying the same thing: x86 architecture is fundamentally flawed! Intel cannot compete!

The problem was huge amounts of x86 code, with many applications partially written in assembly language or, at least, written around the quirks of 8086. Even Intel's own attempt at a next-generation, non-x86, RISC-y chip, the Itanium, couldn't compete.

Intel actually solved the x86's problems partly by embracing RISC concepts: the Pentium Pro and later are, roughly speaking, a RISC processor wrapped up in hardware that translates x86 to the internal RISC instructions.

The new RISC architectures are the wave of the future!

They weren't wrong: ARM chips now greatly outnumber Intel - the bottom has fallen out of the PC market and the growth area is in (mostly ARM-powered) mobile devices. These have 2 key features: (a) they need low power consumption and (b) they don't need to run Windows (MS have proven that Windows isn't an advantage on a phone). Android and iOS have a huge catalogue of apps.

The future might have a better chance this time around: the "Wintel" x86 monopoly has been broken by the growth of mobile & the cloud, and more powerful processors mean that most non-system software can now be written in high-level, processor-independent language, and frameworks for things like graphics are part of the OS, not the application. A lot of modern applications could be converted to ARM just by recompiling - and developers have half-a-mind on iOS/Android versions anyway.

Linux is also a major player (and runs just fine on most RISC chips - again, many applications just recompile or need relatively minor tweaks - most of the big Open Source projects already support ARM). There's interest in ARM servers, too.

Ok, there are still some people who rely on being able to run Windows and I don't think the high-end pro graphics/video people will be going ARM very soon (but then, are they still using Macs?) but its a lot less of an issue than it was 10 years ago. For non-power users - you'd lose the ability to run Windows, but gain the ability to run iOS Apps alongside OS X ones...

The big advantage for Apple, though, is that they could use bespoke A-series processors in everything, made to their specification - no more waiting for Intel to release the right chip.

I could imagine a situation where "Mac whatever Pro" machines were Intel and non-Pro Macs were ARM. Not holding my breath, though.[/QUOTE]

 

[danckwerts]

This story has been going on for years. It might even be a decade. Microsoft has Windows running on ARM processors so it can't be that hard for apple to port Mac OS to ARM. Even Intel might not be so upset about it now they're planning to manufacture ARM chips.

 

[jpietrzak8]

(Actually, no, it was Intel processors that were getting hot: RISC processors were cool!)

Indeed. Although technological advances eventually reversed that; the last RISC chip I remember that was a direct competitor to the x86 line was the DEC Alpha, and DEC had to overclock the poor thing so high to make it competitive that you could use it as a space heater.

The problem was huge amounts of x86 code, with many applications partially written in assembly language or, at least, written around the quirks of 8086. Even Intel's own attempt at a next-generation, non-x86, RISC-y chip, the Itanium, couldn't compete.

Intel actually solved the x86's problems partly by embracing RISC concepts: the Pentium Pro and later are, roughly speaking, a RISC processor wrapped up in hardware that translates x86 to the internal RISC instructions.

Well yeah, that's the thing here. The entire value of RISC was that it allowed for simpler compiler construction, which (in theory) would lead to more efficient code. There were ultimately two problems with this theory: (1) compilers have gotten a whole lot smarter over the last few years, and (2) as Intel showed it's rather easy to implement a complex and powerful instruction set on top of a relatively simple microarchitecture framework.

When you come right down to it, a CPU instruction that allows you to perform a complex task with a single command will potentially be more useful than a CPU that requires you to fire off three simple commands to perform the same task. The problem, of course, is whether or not you can take advantage of that more complex command in your own code; and as compiler technology has advanced, it has become easier and easier to take advantage of complex CPU instructions. As such, the advantage of RISC has receded. Today, the biggest advantage of RISC concepts are entirely hardware-based. (I mean heck, just look at the proliferation of SSE instructions in the x86 architecture, allowing for more efficient graphics processing; someone writing a graphics-intensive program has a much easier time doing it on a modern x86 CPU than on a RISC CPU.)

Another advantage of CISC architecture: smaller executable binaries. The ability to use a single complex command in place of a sequence of tiny commands means less memory used storing those commands, which has many advantages: less disk/solid state storage usage, shorter application loading times, less RAM used, less cache used... There are indeed many subtle costs involved when you try to directly use a reduced instruction set.

 

[ActionableMango]

OP's article says "custom chips", not "custom ARM chips".

I don't know if you guys are following AMD at all, but they've become custom SoC designers and could create custom chips for Apple that are still x86.

I'm far more ready to believe that if Apple is going to switch away from Intel to some sort of custom chip, that the chip would be a custom x86 SoC from AMD, as described here:
http://www.bitsandchips.it/52-english-news/6183-apple-could-use-custom-x86-soc-made-by-amd

 

[Crosscreek]

OP's article says custom chips, not custom ARM chips.

I don't know if you guys are following AMD at all, but they've become custom SoC designers and could create custom chips for Apple that are still x86.

I'm far more ready to believe that if Apple is going to switch away from Intel to some sort of custom chip, that the chip would be a custom x86 chip from AMD, as described here:
http://www.bitsandchips.it/52-english-news/6183-apple-could-use-custom-x86-soc-made-by-amd

I mentioned a Zen Mini in one of these threads.....no one responded so I just dropped it.

We haven't really seen to much on AMD SoC yet and I don't think anyone here has paid to much attention to what going on with AMD.

Seems like a possibility to me.

 

[rGiskard]

(Actually, no, it was Intel processors that were getting hot: RISC processors were cool!)
The big advantage for Apple, though, is that they could use bespoke A-series processors in everything, made to their specification - no more waiting for Intel to release the right chip.

I could imagine a situation where "Mac whatever Pro" machines were Intel and non-Pro Macs were ARM. Not holding my breath, though.

[/QUOTE]

But Apple don't wait on Intel. Most of the Mac lineup now uses CPUs that are a generation old. If anything, I expect that Apple will choose to EOL the "pro" Macs rather than fragment the Mac lineup. They obviously hate developing products for serious computer users so why not go all the way and

Come to think of it, maybe they're now prepping Mac users for such a move. Leave the Pro macs to stagnate and gimp the lower Macs so the shift to ARM won't seem to limit functionality? Not sure I believe it but it could explain their recent moves to abandon pro users.

 

[oneMadRssn]

I think the main thing that held back Macs in the PowerPC days is the lack of software. For a variety of reasons, it was inefficient for developers to make apps for both x86 and PowerPC architectures, and being that x86 was vastly more popular, most chose that path. It wasn't even a Windows vs OS X decision, as when Apple switched to x86, suddenly a world of open source linux apps were made native to OS X. I believe this, in conjunction with being able to emulate Windows apps easily and efficiently, is what tipped Macs over the edge and got us to where we are today.

While ARM-based Macs may certainly exist, and may even be superior to Intel Macs on paper, I think it would be essentially adding back the same old PowerPC handicap again.