With the rumoured name change of OS X to macOS, I was reading between the lines and maybe thought that Apple was going to announce that the new macOS would be running on A9/A10 processors. Now I know that as they are in iPads and iPhones they can't match Intel Core performance but why not stick 2 or 4 of them in the iMacs, Macbooks, etc. Is this even possible? Would the performance increase to match Core i7? How does the GPU capability of the A9 compare with Intel's integrated graphics, for example. I'm not up on such things.
Will Apple ever use their "A" processors in their computers?
- Thread starter ic1male
- Start date
The main problem is that the A processors use the ARM instruction set and the Intel chips use the x86 instruction set. OSX/MacOS itself plus all third party applications would need to be completely rewritten for use on the A series chips.
Its not something thats going to happen anytime soon.
Its not something thats going to happen anytime soon.
Whilst I'm sure Apple could if they wanted (given that they have had enough experience with architecture changes: 68k->PPC->x86 and have previously done software translation for both) - Until there is a real need e.g. to get to the unified ideal of same OS on phone/tablet/laptop/desktop (which hasn't exactly worked that well for Microsoft), then I don't think they will.
https://en.wikipedia.org/wiki/Rosetta_(software)
https://en.wikipedia.org/wiki/Mac_68k_emulator
https://en.wikipedia.org/wiki/Rosetta_(software)
https://en.wikipedia.org/wiki/Mac_68k_emulator
Deleted member 138126
Deleted member 138126
Nah, too much effort for little gain. The Intel M processors are pretty competitive power-efficiency-wise. There may be Macs running iOS though, in which case, yes, they would run on Arm processors.
This. Not gonna happen. Too many programs rely on x86 architecture that can't just be recompiled on a whim by their authors.
Associate
- Joined
- 27 Jan 2014
- Posts
- 892
Can't they just run in the way PowerPC apps ran when they moved to intel? Well, not the same way (Rosetta) but the equivalent.
Deleted member 138126
Deleted member 138126
They *could*, but why would anyone want to? The move from PowerPC to Intel was necessary because PowerPC was a dead technology. Intel is far from dead, and there are plenty of x86 low-power alternatives right now that are "good enough".
Soldato
- Joined
- 4 Oct 2008
- Posts
- 6,693
- Location
- London
Firstly you can't compare it at all too the Power PC -> Changes.
* ARM doesn't have anything right now that comes even close too the higher end, so it would mean OSX has too run on 2 platforms for the foreseeable future.
* Transition apps would be much harder this time, there is a much larger user base, way more OSX apps.
* Intel isn't dead, right now they still have the best fabs (no one is even close), they can power things like the Macbook very well without a fan, and they have only just started focusing on power usage is recent years. It seems like a very difficult and expensive change for small gains in the long run. Intel delays are caused by the difficult of getting too smaller and smaller nodes, ARM is mostly on 16/14nm so they are years behind intel.
Also I'm pretty sure ARM is yet too support things like ThunderBolt, unless that changes it would mean things like Thunderbolt displays are useless over night for a portion of Macs Apple makes. Honestly don't see the gain.
* ARM doesn't have anything right now that comes even close too the higher end, so it would mean OSX has too run on 2 platforms for the foreseeable future.
* Transition apps would be much harder this time, there is a much larger user base, way more OSX apps.
* Intel isn't dead, right now they still have the best fabs (no one is even close), they can power things like the Macbook very well without a fan, and they have only just started focusing on power usage is recent years. It seems like a very difficult and expensive change for small gains in the long run. Intel delays are caused by the difficult of getting too smaller and smaller nodes, ARM is mostly on 16/14nm so they are years behind intel.
Also I'm pretty sure ARM is yet too support things like ThunderBolt, unless that changes it would mean things like Thunderbolt displays are useless over night for a portion of Macs Apple makes. Honestly don't see the gain.
Apple is usually quite pragmatic in it's choice of components. What would they benefit from switching from x86 to ARM?
Macbook runs on Intel's latest and greatest x86 SoC and clocks pretty respectable battery life from its 41.4Wh battery.
Ipad Pro 12.9 has 38.5Wh battery and while it arguably gets a bit better battery life figures, the difference is not very big and depends on workloads and tests.
Here is a small comparison of CPU benchmarks:
http://www.anandtech.com/show/9766/the-apple-ipad-pro-review/4
So A9X is not faster, but somewhat slower than skylake.
So again, what would Apple gain by switching from Intel to their own SOC's? At the moment the result is simple: Reduced price and therefore higher margins / or ability to price their laptops lower. Intel charges a lot of money for their high-end Core M's. Performance wise A9X can't match Skylake Core M or especially Core i5's and i7's in bigger laptops.
Sure, A10(x) will again be a bit better than A9x, but intel will keep progressing as well. We can be sure that Apples huge progress in performance of A-series CPU's and GPU's is going to slow down considerably in future. Low hanging fruits have been collected.
I'm willing to bet a decent amount of money that this transition won't happen any time soon. But "ever" is a very long time. Going 5 years in future and who knows. Maybe xcode is developed enough that dev's can just hit "compile for mac" button and both, ARM and x86 binaries are created. Metal is supported on both so GPU's are going to be fine etc.
Macbook runs on Intel's latest and greatest x86 SoC and clocks pretty respectable battery life from its 41.4Wh battery.
Ipad Pro 12.9 has 38.5Wh battery and while it arguably gets a bit better battery life figures, the difference is not very big and depends on workloads and tests.
Here is a small comparison of CPU benchmarks:
http://www.anandtech.com/show/9766/the-apple-ipad-pro-review/4
So A9X is not faster, but somewhat slower than skylake.
So again, what would Apple gain by switching from Intel to their own SOC's? At the moment the result is simple: Reduced price and therefore higher margins / or ability to price their laptops lower. Intel charges a lot of money for their high-end Core M's. Performance wise A9X can't match Skylake Core M or especially Core i5's and i7's in bigger laptops.
Sure, A10(x) will again be a bit better than A9x, but intel will keep progressing as well. We can be sure that Apples huge progress in performance of A-series CPU's and GPU's is going to slow down considerably in future. Low hanging fruits have been collected.
I'm willing to bet a decent amount of money that this transition won't happen any time soon. But "ever" is a very long time. Going 5 years in future and who knows. Maybe xcode is developed enough that dev's can just hit "compile for mac" button and both, ARM and x86 binaries are created. Metal is supported on both so GPU's are going to be fine etc.
Well MS already have ARM versions of their products (for their Surface line) so that would be easy for them. Adobe do code for ARM so probably would have the experience to recode PS.
But as already said, why would Apple do it? Unless there were significant space, performance or battery reasons then I don't see why they would. Unless the business case for moving even more stuff "in house" is strong enough.