That's what he meant by a single OS. I didn't get that. I agree with you that Apple could put Mac OS X on an ARM64 system and sell a profitable Macs (or Mac-like devices) that are still undercutting the competition while being very fast and versatile machines.
This is an interesting discussion. There are reasons why Apple can't put OS X on an A7, of course. But that doesn't mean they won't be able to with a more advanced device. The A8 is expected to have four cores. That will almost double multi core performance. And I wonder how much per core performance will increase. I don't see them as being able to double per core performance each year anymore, particularly if they use the area for two more cores. But even if they increased it by 25% this year, multicore performance would rise by an additional 50%. For example, say per core performance of the A7 is 100, and we have perfect correspondence with multi core, impossible, but it's very close, the multicore performance is 200. Now, if per core performance is moved to just 125, and we have four cores, then multicore performance jumps to 500. That's Huge!
But Apple could have some other surprises under their sleeves. One of the biggest problems with running an OS under a different chip is that some functions aren't directly duplicatable (is that a word?). So they either need to be done in software, which can be anywhere from two to a thousand times slower, or use extra hardware steps, also very slow. Both cause the dramatic slowdowns in emulation. That 's why Windows was so slow on the PPC. It's estimated to get equal speed across lines in that way, the chip on which the emulation is being done must be roughly five times as fast (or powerful, depending on how you want to describe it) as the chip being emulated. So while for several years, the PPC was over 50% faster than the Pentium, Windows ran at about 20% (at best!) of the speed it ran on an inexpensive PC.
So there's the rub. Right now, there is no way at all an ARM chip could run OS X reasonably.
But there actually is a way, should Apple choose to take it. See, before, they had some say over the direction of PPC, but mostly through writing firmware and micro architecture for the chips. They didn't control the actual direction other than indirectly. But now...
From what I know, there are functions in x86 that are legally allowed to be duplicated in hardware. These are individual functions, not the chip design for which a company would need to get the non existant license for. A number of these functions are the same ones that are the most difficult to emulate, and so take up most of that processing time, and so also slow the emulation down the most. I hope everyone is getting what I'm saying.
Apple could, if it wanted to, add those functions to its ARM designs. They wouldn't necessarally be needed for normal ARM operation, though they could be, if they give some advantage. But, if Apple chose to use that chip for OS X, well, these would cut that emulation hole down considerably. Possibly enough so that next years' chip, assuming that they will do this, could run OS X at the speed of an ultra low power x86.
In addition, if Apple was willing to have a seperate line of chips from iOS, they could account for the higher cooling in a notebook and run them at a somewhat higher speed, adding to the power of the chip. Also, considering just how cheap Apple's SoC's are, they could even use two in the machine, with a bus similar to what intel uses for the Xeon, and still have a price advantage.
So with some major changes, I can see this happening. But there need to be major changes, or performance will suffer unacceptably.
"Can Apple's A* chips (A7 like chips) power Mac Notebooks like MBA, MBP etc.?"
It's possible but it will be way slower. If you think the A7 compares equally to an Intel chip your mistaken. It would take you back in time about 4-5 years as far as speed. Also, anyone who uses a VM like Parallels to run Windows is out of luck. So no Windows Quickbooks or Quicken. On the plus side it will be more energy efficient so longer battery life.
I really don't see Apple introducing a brand new Macbook line running OSX. It seems to be they would want to propagate and promote iOS to the mainstream more than OSX. The most feasible product I see is the following, which I believe deserves to exist and can fill a niche:
- A8 based machine in a clamshell form factor
- iOS8 based, with some UI elements that cater more to desktop use
- 12" touchscreen, almost edge to edge
- Physical keyboard
- No ports beyond power and headphone jack- same as iOS devices
- no fans, no moving parts
- A touchpad that can actually control iOS
- Ultralight, ultrathin, ultra long battery life.
I think this would actually be the perfect machine for many. iOS can then be further legitimized for desktop applications, and for most it would be enough. Integration of touchpad/mouse functionality is a must for this form factor. Dont see why it cant be done- especially with the further maturing of iOS8 and iCloud.
Samsung and HP are selling Google Chromebooks using the Exynos chips. They work OK but they are definitely slower than the Intel Haswell chips. I would like to see what Apple could do with an A7 or A8 chip in a laptop machine. The big difference would be that OS X is much larger than Chrome OS.
Only about an extra 3% in performance according to AnandTech.
It's not much different from the prior Ivy --> Ivy chips. I would call the price adjustments much bigger news as they place a 13" machine below $1000. Last time they had one in that spot was the white macbook IIRC.
I prefer iQuickMark - it covers not just CPU, but also other important for performance aspects like RAM and Flash. The latter is quite critical for performance.
Comments
This is an interesting discussion. There are reasons why Apple can't put OS X on an A7, of course. But that doesn't mean they won't be able to with a more advanced device. The A8 is expected to have four cores. That will almost double multi core performance. And I wonder how much per core performance will increase. I don't see them as being able to double per core performance each year anymore, particularly if they use the area for two more cores. But even if they increased it by 25% this year, multicore performance would rise by an additional 50%. For example, say per core performance of the A7 is 100, and we have perfect correspondence with multi core, impossible, but it's very close, the multicore performance is 200. Now, if per core performance is moved to just 125, and we have four cores, then multicore performance jumps to 500. That's Huge!
But Apple could have some other surprises under their sleeves. One of the biggest problems with running an OS under a different chip is that some functions aren't directly duplicatable (is that a word?). So they either need to be done in software, which can be anywhere from two to a thousand times slower, or use extra hardware steps, also very slow. Both cause the dramatic slowdowns in emulation. That 's why Windows was so slow on the PPC. It's estimated to get equal speed across lines in that way, the chip on which the emulation is being done must be roughly five times as fast (or powerful, depending on how you want to describe it) as the chip being emulated. So while for several years, the PPC was over 50% faster than the Pentium, Windows ran at about 20% (at best!) of the speed it ran on an inexpensive PC.
So there's the rub. Right now, there is no way at all an ARM chip could run OS X reasonably.
But there actually is a way, should Apple choose to take it. See, before, they had some say over the direction of PPC, but mostly through writing firmware and micro architecture for the chips. They didn't control the actual direction other than indirectly. But now...
From what I know, there are functions in x86 that are legally allowed to be duplicated in hardware. These are individual functions, not the chip design for which a company would need to get the non existant license for. A number of these functions are the same ones that are the most difficult to emulate, and so take up most of that processing time, and so also slow the emulation down the most. I hope everyone is getting what I'm saying.
Apple could, if it wanted to, add those functions to its ARM designs. They wouldn't necessarally be needed for normal ARM operation, though they could be, if they give some advantage. But, if Apple chose to use that chip for OS X, well, these would cut that emulation hole down considerably. Possibly enough so that next years' chip, assuming that they will do this, could run OS X at the speed of an ultra low power x86.
In addition, if Apple was willing to have a seperate line of chips from iOS, they could account for the higher cooling in a notebook and run them at a somewhat higher speed, adding to the power of the chip. Also, considering just how cheap Apple's SoC's are, they could even use two in the machine, with a bus similar to what intel uses for the Xeon, and still have a price advantage.
So with some major changes, I can see this happening. But there need to be major changes, or performance will suffer unacceptably.
http://69.165.170.78/images/GeekBench_Mid2012_MPA.jpg
It's possible but it will be way slower. If you think the A7 compares equally to an Intel chip your mistaken. It would take you back in time about 4-5 years as far as speed. Also, anyone who uses a VM like Parallels to run Windows is out of luck. So no Windows Quickbooks or Quicken. On the plus side it will be more energy efficient so longer battery life.
I really don't see Apple introducing a brand new Macbook line running OSX. It seems to be they would want to propagate and promote iOS to the mainstream more than OSX. The most feasible product I see is the following, which I believe deserves to exist and can fill a niche:
- A8 based machine in a clamshell form factor
- iOS8 based, with some UI elements that cater more to desktop use
- 12" touchscreen, almost edge to edge
- Physical keyboard
- No ports beyond power and headphone jack- same as iOS devices
- no fans, no moving parts
- A touchpad that can actually control iOS
- Ultralight, ultrathin, ultra long battery life.
I think this would actually be the perfect machine for many. iOS can then be further legitimized for desktop applications, and for most it would be enough. Integration of touchpad/mouse functionality is a must for this form factor. Dont see why it cant be done- especially with the further maturing of iOS8 and iCloud.
Samsung and HP are selling Google Chromebooks using the Exynos chips. They work OK but they are definitely slower than the Intel Haswell chips. I would like to see what Apple could do with an A7 or A8 chip in a laptop machine. The big difference would be that OS X is much larger than Chrome OS.
Only about an extra 3% in performance according to AnandTech.
It's not much different from the prior Ivy --> Ivy chips. I would call the price adjustments much bigger news as they place a 13" machine below $1000. Last time they had one in that spot was the white macbook IIRC.
There are no change in inputs/outputs for the new Airs. Oh, and they have what most people need.
cool
and yea...I figure if you need to connect a lot of things to your laptop you need adapters or whatever anyways.
I prefer iQuickMark - it covers not just CPU, but also other important for performance aspects like RAM and Flash. The latter is quite critical for performance.