Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ARM chips have additional performance penalties compared to x86 that these benchmarks do not take into account. x86 has far superior memory and storage subsystem performance.

ARM chips in current mobile phones and tablets have those penalties, which slow them down in benchmarks.

Apple designed the memory controllers for many of the PowerPC Macs, and they were no slouch in memory performance (given the process technology at that time). So there's no reason that a 64-bit ARM chip specifically (re)designed for a laptop by Apple won't have a laptop class memory system and laptop-sized heat sink, thus massively increasing system performance as well as top-end clock speed over the mobile phone chips. There's likely nothing in the logic design of the arm64 CPU cores in the A8X to prevent that, and Apple has enough engineers to tweak and tune the physical design and layout for a hotter process....

...assuming they actually want to do such a product, and not just put a lot more heat on their current supplier, Intel, to keep up with the roadmap.
 
My 2 cents here

1. Cent moving to arm does lock down the mac further, look a win 8 on arm (or windows rt if you will) only apps in the windows App Store are allowed to be installed same thing would happen with the mac. No side loading. Also the amount who have banged on about windows 8 running on a mac with arm, you do all know that windows rt failed and sucked harder than normal windows 8 so much so that there was a surface 2 pro and rt but there is no surface 3 rt. That should tell you everything MS thinks of windows on arm hell not even HP and asus selling windows RT devices they did by the way. And why did they stop selling them? because no body brought them even though they came with office built in for free.

2nd cent PPC to intel was a lot easier than this, why?? Because I don't know if any one knows this but apparently not from most peoples post about this, but the other 95% of the pc market was on X86 so adobe, MS and anybody you can think of had X86 basic code and all they needed to do was to re compile them for the mac. There is no full fat software available on ARM on any platform, it's all baby software, just look at windows RT on arm and what a failure that has been. Also notice how MS hasn't spoken about RT since the surface 2 from a product and platform for developers stand point, so if you all want to run windows rt on your arm mac go ahead because it can't do anything.
 
I don't see any big trouble from the switch for existing users - as long as they don't change the API (most apps would just need to be recompiled in Xcode) ...

however, the main reason I switched to Mac was that they went 'Intel' - it was an easy decision, I wasn't sure if MacOS is right for me, but I could always have turned around and installed windows on it. For a while I ran parallels since I didn't get every windows app right away for the mac. But that was years ago and I never looked back, I just love the Mac.

but even today, there are plenty of windows users that might consider a mac, knowing that they have a backup plan (installing windows) or at least run a fast and stable parallels/vmWare/....

I see the advantages of arm - but for growing the user base intel might be the better option.

May they go both ways - that you can buy both intel and arm and all the apps will be big fat universal apps running on both platforms (maybe stripped down during installation)
 
Yep lets go on another crazy switch for NO REAL REASON. The A8X is not powerful enough to run a desktop class operating system.

The only thing switch would do would mean having to recompile things and confuse people over what platform they are buying as it will effect what they are able to do on their Mac.

Maintaining two versions of OSX would also be a logistics nightmare. It would also take Apple around 2-3years to get OSX right for the platform, as it wasn't till SL that we had a proper Intel OS that performed as it should.

I've got no doubt that Apple is testing ARM based Macs, and so it should be. It would be testing a heck of a lot of things. Heck they're probably tested Intel Tablets and iPhones, but that doesn't make it a good idea.

I never said it was a good idea. The only thing I said is that CPU power isn't a reason this wouldn't work ( A8X is a very powerful SoC, and it is powerful enough for a typical customer and it'd probably be enough for some "pro" stuff). Also, considering that an Apple AX chip in a future Mac would be at least twice as fast (probably much more due to larger battery or no battery problems at all) than A8X, I see no reason for that being a problem.

If we're talking about hardcore pro stuff, then you might have a point.
 
Apple is far bigger now than they were back then - even as the PC market shrinks Macs stay pretty consistent.

I think they have something like 12% of the PC market.

12% is very small. It will be significantly smaller if they keep going in the proprietary path they are heading down.
 
"based on the assumption that Apple's self-developed AP performs at a level between Intel's Atom and Core i3 and is good enough for Mac."

Apple might think that something less than an i3 is good enough. I certainly would not.
 
You're correct the CPU has noting to do with the OS being locked down, but is hard to see apple not doing this if they go the ARM approach.

If Apple is going to close OS X they are going to do it with or without a move to their own CPUs.

I don't really see that they are even moving in that direction. If they were, I don't think they would have let their Mac App Store languish.


Geekbench is a terrible cross architecture test. A better one would be application tests that can be performed on both platforms. Let go for a pretty basic one, web browsing. As we can see from the below image, the i3 still carries a significant performance advantage, even within a thermally compromised chassis like the Surface 3.

Image

ARM chips have additional performance penalties compared to x86 that these benchmarks do not take into account. x86 has far superior memory and storage subsystem performance.

Geekbench is a limited benchmark... that is to say, *it is a benchmark*.
I would agree that it's a terrible cross architecture test. But it's also one of the best available.

At least it's actually designed to be a cross-platform benchmark... the browser tests are not (though they aren't an unreasonable proxy due to the strong competition for performance in browsers these days).

Anyway, don't your benchmarks show the iPad Air 2 getting in the ballpark of the performance of the Surface Pro 3 with i3 and about half that of one with an i5?

And regarding your last part, Geekbench does test memory subsystem performance. (And it's not like the browser-based benchmarks factor that out somehow.)
 
If I remember correctly, MacOS X 10.4 Tiger was the first OS, that had a PPC and an Intel version.

So who knows, if Apple doesn't have already a version of Yosemite made for ARM?

I can eventually see, that the Macbook and Mac mini line (even the iMac) can be ARM based in the next years.

About the Mac Pro ... who needs that? How much money does Apple make from the "pro" section? If they would, they didn't drop Aperture altogether. So I can see a future, when there will be no more Mac Pro, nor any of these so called "Pro Apps".

When they announced Intel integration, Jobs confirmed the rumor that they had been secretly developing all OS X versions with x86 processors. It probably didn't run well, but it was there. I bet they're doing the same thing with ARM processors already.
 
Why not both

Here is my pure speculation. Why not use both Intel and ARM together? Just like they do with GPUs (in some models) in the past? ARM for basic needs like web browsing, iTunes and such, and Intel only when running x86 (or 64) binaries or during virtualization. That could potentially save a lot of power, if Intel chip is left at sleep, or could leave the Intel cores for more power hungry apps, reducing task switching overheads and increasing perceived speed of processing for those apps.

I am not an OS or logic design expert. Just speculating here.
 
Here is my pure speculation. Why not use both Intel and ARM together? Just like they do with GPUs (in some models) in the past? ARM for basic needs like web browsing, iTunes and such, and Intel only when running x86 (or 64) binaries or during virtualization. That could potentially save a lot of power, if Intel chip is left at sleep, or could leave the Intel cores for more power hungry apps, reducing task switching overheads and increasing perceived speed of processing for those apps.

I am not an OS or logic design expert. Just speculating here.

Yap - in the current state iTunes is in it definitely needs its own dedicated processor.
 
I'm sorry but once they stop making Intel mac, is the day I stop buying them ...

Which might be a great trade-off for Apple. Some percentage of current Mac owners would stop buying them. Some percentage of iPad and iPhone users looking for a new laptop would start buying this hypothetical new product line (because an ARM based Mac could be priced more aggressively than the current ones, even while keeping Apple's nice profit margins). The latter group might be larger, thus increasing total Mac unit sales, without you.
 
And regarding your last part, Geekbench does test memory subsystem performance. (And it's not like the browser-based benchmarks factor that out somehow.)

With Javascript tests you are going to be sitting in cache most of the time.

Oh and, ARM server (no power and heat constraints) memory performance: http://www.anandtech.com/show/6757/calxedas-arm-server-tested/6 Then take a look at something that stresses memory and storage: http://www.anandtech.com/show/6757/calxedas-arm-server-tested/8 ouch! I know that is a Xeon (Intel do share memory controller design across their mainline x86 chips), but it still puts overall ARM vs x86 performance into perspective.

As for the people naively thinking you can just bump up the Mhz once in a laptop/desktop chassis. Intel made the same mistake in thinking a decade ago, we got the P4.
 
Last edited:
I'm not I understand the concern with not being able to use Windows for work if Apple switches to ARM.

With remote desktop apps who can't connect with their work PC and use their Mac? I have done it with my iPad at my former employer (I'm now retired).

Regardless, I have a hard time believing that Apple would abandon high end processing altogether and only offer relatively low powered processors.

I am willing to bet Apple's long term business strategy is pretty sound so why worry?
 
Such a machine might be worth it for power consumption alone

A 100W machine running 24/7 for 5 years at 14 cents a KWHr eats up over $600 of electricity.
 
I may have missed the boat on this conversation reading everyone's views. But I feel like we are missing a major point:

I won't dispute the change in architecture would be a royal pain in the a$$ - and losing Windows compatibility would be a major hit...but realistically people seem to be saying *IF* this happens it will be in a decade? Then why is everyone trying to compare the A8X to the Atom/i3/i5? The iPhone isn't even 10 years old and here we are trying to speculate the future of ARM vs x86 based on the performance of Apple's "fifth" chip in 4 years?

The future of computing is always slimmer and smaller (I know many here don't like that about Apple, and I am even frustrated at times...but I think it's true nonetheless) - if ARM lends itself to that skill set more than the current regime - I wouldn't be surprised to see Apple part ways.

I'm just saying it's incredibly hard to predict the next computing breakthrough. I would be surprised if Intel couldn't find a way to make x86 competitive with ARM in all categories in the next TEN years...but I'm also the guy who believed Steve Job's first picture of the iPhone (you know the one with the rotary dial?) ;)
 
I still dont get why people want to see ARM based Macs.
Or the prediction at the end of the decade.
Why? I only see headaches with this idea.

What headaches? Either B-U-Y ARM-compatible versions of your apps,
or you're out on your ass. What's not to understand? lol
 
I still dont get why people want to see ARM based Macs. Or the prediction at the end of the decade. Why? I only see headaches with this idea.

I'm pretty sure Apple doesn't pay 300$USD or more for its own Ax CPUs. If it means a much lower price for the MacBook Air and the Mac mini, as long as my software keeps running, I don't have a problem with that.

Most of the software I use is made by Apple, otherwise all I really need is Photoshop, TextWrangler, an FTP program of some sort and Sketchup.
 
Last edited:
I'm pretty skeptical. I'd expect that Apple more than anyone else would have reservations about switching to desktop architecture no one else is using. They've been burned here before. Developing high-performance microprocessors is very expensive. The market can *barely* sustain two suppliers as it is, as AMD continues to flounder. Would Apple commit to, essentially permanently, contributing substantial resources to keep up with Intel? Consider that this involves heavy research into process improvements and physical properties well outside Apple's specialties.

Making high volume chips for mobile devices that build on existing processes? Absolutely.

Investing the resources in researching and shrinking fabrication techniques necessary to compete with companies who focus solely on this kind of thing? And the enormously-expensive fabs necessary to support them? *And* deliver in the volume required to support their product lines. I'm not convinced.
 
...but I'm also the guy who believed Steve Job's first picture of the iPhone (you know the one with the rotary dial?) ;)

You mean ... ?

knIPHONE_wideweb__470x303,0.jpg
 
Isn't Kuo female? As in, not he? Are we sure she is a male? As I do know that women have the same exact name. Anyone want to weigh in here?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.