It would mean laptops (mostly) with tailored made SoC technology.
Look at it this way a SoC is effectively the equivalent of a printed circuit board from the 80's to 90's time frame. It is where engineers now glue their unique designs together. Instead of designing a printed circuit board and a few custome chips to realize most of your platform, you instead design a SoC much like Apple has with the A series chips.
Now why would Apple want to do this for a Laptop? It allows for exactly what they need and want on the SoC instead of settling for Intels generic solutions.
That is all right, it is actually a good question. From the standpoint of people want to see this happen I think it boils down to these issues:
We want to see Intel under constant pressure to improve their hardware. In the past AMD was a big factor here, today not so much.
ARM might be a better fit for ultra compact laptops like the new Mac Book. The thing is we really don't know how well the new Mac Book will perform due to thermal issues common to Intel hardware. Apples ARM designs often run flat out while keeping performance stready.
ARM would allow far more cores on chip than can be had with i86 which in some workloads would be ideal. Imagine if you will a six core SoC, A10 if you will, that only runs at 5 watts. This might sound extreme to some but the ARM cores aren't massive and they can be improved. If you look at the photos of the current Apple implementation you will see that six cores would be easy space wise with a process shrink.
Remember Apple hasn't even implemented a FinFet type processor yet. Doing so could allow them to produce a decent machine with a 2 watt SoC.
It puts Apple in the drivers seat.
Lastly a big reason is cost. Intel gets a really good price on its laptop chips. Going ARM, leveraging IOS hardware development costs might allow Apple to trim $300 of the cost of a machine. This especially if they can implement stuff (on the SoC) that requires third party chips on Intel based machines.
And the MacBook has been and still is "very laptop like in its construction."
Like what? No touchscreen, a physical keyboard, a clamshell, and a windowed OS still make it a notebook computer.
You seem to have a problem with what many of us see as obvious. The Mac Book pulls concepts from the tablet space, that doesn't make it a tablet but you can't deny that it draws upon the tablet world. Consider the minimalist mother board, the single port, the design optimized batteries, thinnest and other features that where first seen in the tablet world.
There is absolutely nothing tablet-esque about that. It's not a "transformer" laptop/tablet device.
Everything about this machine tells me they applied what they learned in building tablets to it. It is still a laptop but one obviously derived from Apples tablet experiences. I'm really not sure how you can mis this. A laptop doesn't have to look like any sort of tablet to embody many tablet concepts.
Being thinner, being fanless, and losing ports do not a tablet make. If you think those make it tablet-esque then why not the MBA arriving the year after the iPhone and well after the iPod make it smartphone-esque or PMP-esque simply because it copies some of the simplicity to its design?
At the time the MBA arrived it was nothing more than a compacted laptop and didn't have tablets to be compared to. In the case of this Mac Book it is very rational to compare it to a tablet. The way I see it, Mac Book is a solution to the "tablet with keyboard" niche. That doesn't mean that it is a tablet just that Apple sees the design as a better solution that some of the hybrid stuff on the market.
If Apple does move to an ARM-based Mac or Mac-like "PC" at the low end that also won't make it just like a tablet, although I'm sure there will be some that will claim it is, jus as now there are people claiming that Core-M is just like Atom.
Well comparing Core M to Atom is just ignorance which you have rightfully pointed out.
As for Apple and ARM it is the operating system that defines the machine. Mac OS is what makes the Mac Book a Mac.
You seem to have a problem with what many of us see as obvious. The Mac Book pulls concepts from the tablet space, that doesn't make it a tablet but you can't deny that it draws upon the tablet world. Consider the minimalist mother board, the single port, the design optimized batteries, thinnest and other features that where first seen in the tablet world.
Everything about this machine tells me they applied what they learned in building tablets to it. It is still a laptop but one obviously derived from Apples tablet experiences. I'm really not sure how you can mis this. A laptop doesn't have to look like any sort of tablet to embody many tablet concepts.
At the time the MBA arrived it was nothing more than a compacted laptop and didn't have tablets to be compared to. In the case of this Mac Book it is very rational to compare it to a tablet. The way I see it, Mac Book is a solution to the "tablet with keyboard" niche. That doesn't mean that it is a tablet just that Apple sees the design as a better solution that some of the hybrid stuff on the market.
Well comparing Core M to Atom is just ignorance which you have rightfully pointed out.
As for Apple and ARM it is the operating system that defines the machine. Mac OS is what makes the Mac Book a Mac.
And? The iPad pulled concepts from the iPhone, which pulled concepts from their conceptual tablet which was based on off expertise learned from the iPod and Mac? That DOES NOT make the MacBook a tablet, tablet-like, or tablet-esque. This is simply the natural progression of notebook computers. We can look to the original MacBook Air, which predates the iPad by many years, to see that this shrinking logic board, non-removable battery, single USB port, etc. were enacted. It's silly to say that because we finally have a USB standard that can be used for supplying power to a notebook that it's becoming an iPad. By that logic, Apple's inclusion of revamped frameworks and APIs in iOS that have made Mac OS X more efficient is the Mac becoming an iPhone. That's simply ridiculous.
The way I see it, Mac Book is a solution to the "tablet with keyboard" niche.
A non-convertible notebook with a windowed OS, mouse pointer, and no touchscreen is NOT a tablet, with or without a keyboard. An actual tablet computer or convertible, with a keyboard is the "tablet with keyboard" niche.
It is more than a notch up. You need to do some research here, this is a Broadwell based processor.
It's a smaller die on broadwell and low power 5w on the M series. It's not out performing the Haswell range, that will come later this year with skylake. If you want a thermally limited lower power cpu than the previous gen Airs then this is for you.
The downside is incompatibility on the software side with the existing software base compiled to x86 and the loss of Windows compatibility via Bootcamp or parallels.
Also, while Intel has shown that they can play down in the <5W TDP range with ARM, ARM has not yet shown they can scale up to Core i7 performance levels.
That Intel has a 4.5W part that performs almost at Core i5 levels is a significant indicator that the window of opportunity for ARM to displace x86 on the MacBooks closed. A few years back there was the thought that if Intel stumbled on the performance per watt category that ARM would massively displace them in the notebook and server markets. With Core-M the notebook transition seems less likely today and with Xeon-D the same can be said of the server market.
The downside is incompatibility on the software side with the existing software base compiled to x86 and the loss of Windows compatibility via Bootcamp or parallels.
Also, while Intel has shown that they can play down in the <5W TDP range with ARM, ARM has not yet shown they can scale up to Core i7 performance levels.
That Intel has a 4.5W part that performs almost at Core i5 levels is a significant indicator that the window of opportunity for ARM to displace x86 on the MacBooks closed. A few years back there was the thought that if Intel stumbled on the performance per watt category that ARM would massively displace them in the notebook and server markets. With Core-M the notebook transition seems less likely today and with Xeon-D the same can be said of the server market.
At least for the near future (5-6 years).
1) Apple has done the transitions several times with great success. I don't that today, with an centralized Mac App Store in place, it will be an issue.
2) Why does it need tone Core i7 performance levels? I'd think it would start on the lower-end of the performance scale. Note that the performance of CPUs is not a driving factor for the average customer.
3) I think Core M performance indicates Apple is much more likely to offer an ARM-based traditional "PC".
It would mean laptops (mostly) with tailored made SoC technology.
Look at it this way a SoC is effectively the equivalent of a printed circuit board from the 80's to 90's time frame. It is where engineers now glue their unique designs together. Instead of designing a printed circuit board and a few custome chips to realize most of your platform, you instead design a SoC much like Apple has with the A series chips.
Now why would Apple want to do this for a Laptop? It allows for exactly what they need and want on the SoC instead of settling for Intels generic solutions.
That is all right, it is actually a good question. From the standpoint of people want to see this happen I think it boils down to these issues:
We want to see Intel under constant pressure to improve their hardware. In the past AMD was a big factor here, today not so much.
ARM might be a better fit for ultra compact laptops like the new Mac Book. The thing is we really don't know how well the new Mac Book will perform due to thermal issues common to Intel hardware. Apples ARM designs often run flat out while keeping performance stready.
ARM would allow far more cores on chip than can be had with i86 which in some workloads would be ideal. Imagine if you will a six core SoC, A10 if you will, that only runs at 5 watts. This might sound extreme to some but the ARM cores aren't massive and they can be improved. If you look at the photos of the current Apple implementation you will see that six cores would be easy space wise with a process shrink.
Remember Apple hasn't even implemented a FinFet type processor yet. Doing so could allow them to produce a decent machine with a 2 watt SoC.
It puts Apple in the drivers seat.
Lastly a big reason is cost. Intel gets a really good price on its laptop chips. Going ARM, leveraging IOS hardware development costs might allow Apple to trim $300 of the cost of a machine. This especially if they can implement stuff (on the SoC) that requires third party chips on Intel based machines.
If multicore ARM was the solution, android would have the market cornered. I thought we learned a LONG time ago, that just adding more cores only works with GPUs?
1) Apple has done the transitions several times with great success. I don't that today, with an centralized Mac App Store in place, it will be an issue.
Neither Bootcamp nor Parallels will efficiently run on ARM. Likewise emulating x86 compatibility with a Rosetta like ability was livable for users because the Core Duo outperformed the G5 by the time the transition happened.
In the Ars benchmarks the Core Duo iMac generally outperformed the G5 iMac using universal binaries.
There was also a very long period where universal binaries did not exist for many apps including many of Apple's own. The existence of the Mac App store does not magically translate into programmer hours required to do even fairly minimal translation, recompile and most importantly retesting.
Quote:
2) Why does it need tone Core i7 performance levels? I'd think it would start on the lower-end of the performance scale. Note that the performance of CPUs is not a driving factor for the average customer.
For the same reason that when Apple transitioned from PowerPC to X86 they did so across the entire computer line as quickly as possible. Software compatibility across the entire line up is hugely important for developers. To leave the MBP, iMac and Mac Pro on x86 while the Macbook is ARM results in the same mess Microsoft ended up with between Windows and Windows RT Surface platforms.
No transition to ARM will occur unless Apple is sure that: a) windows compatibility will be unimportant for business users and b) the top end iMac, MBP and Mac Pro will be faster than the same generation Intel offerings or the heavy professional users will likely jump ship to Windows because Time = Money.
Quote:
3) I think Core M performance indicates Apple is much more likely to offer an ARM-based traditional "PC".
This is a ridiculously silly position. The Core M shows that ARM provides very little advantages over X86 on traditional PC platforms except on price. If Intel was still stuck at 15W+ TDP while ARM could provide similar performance at <5W then there would be a significant case for moving from x86 to ARM.
At 4.5W Intel offers Apple a way forward with fanless designs with long battery life while keeping software compatibility from the thinnest ultra book to the top end Mac Pro.
If anything it indicates that Apple is much more likely to offer an x86-based pro tablet (or convertible) someday running OSX and with a digitizer in addition to multi-touch for creatives. More likely, not likely. Odds are still pretty low but better than zero which is what it was when decent Intel performance was still in the 15W range.
If multicore ARM was the solution, android would have the market cornered. I thought we learned a LONG time ago, that just adding more cores only works with GPUs?
:???: And Apple was foolish to buy PA Semi and think they could ever do anything with ARM compared compared to the incumbents, especially coming out with a 64-bit SoC with a full 64-bit version of iOS years before anyone else. Where is Andrioid's cornering of the AArch64? It doesn't exist because there the benefit isn't there without vertical integration and proper planning. That isn't to say that Apple will offer more cores — who knows, perhaps they'll stay at 2-cores for ever — but I think that such a scenario is unlikely.
And Apple was foolish to buy PA Semi and think they could ever do anything with ARM compared compared to the incumbents, especially coming out with a 64-bit SoC with a full 64-bit version of iOS years before anyone else. Where is Andrioid's cornering of the AArch64? It doesn't exist because there the benefit isn't there without vertical integration and proper planning. That isn't to say that Apple will offer more cores — who knows, perhaps they'll stay at 2-cores for ever — but I think that such a scenario is unlikely.
What I was addressing was his idea of just taking an iPhone SoC and adding some more cores. I mentioned Android because the trend there seems to be to just add more cores, and in fact, Apple did the opposite starting with the A6 SoC.
Why don't if for example a chip foundry has plans for going from say 30 to 22 to 14 nm, why don't they start designing 14, when the 30 is released? Why is it that they need to go down in steps. I presume there is an actual reason and not just the obvious cynical reason.
Comments
It would mean laptops (mostly) with tailored made SoC technology.
Look at it this way a SoC is effectively the equivalent of a printed circuit board from the 80's to 90's time frame. It is where engineers now glue their unique designs together. Instead of designing a printed circuit board and a few custome chips to realize most of your platform, you instead design a SoC much like Apple has with the A series chips.
Now why would Apple want to do this for a Laptop? It allows for exactly what they need and want on the SoC instead of settling for Intels generic solutions.
That is all right, it is actually a good question. From the standpoint of people want to see this happen I think it boils down to these issues:
Excellent. Now it all makes a lot more sense.
Thanks for taking the time to explain.
Well comparing Core M to Atom is just ignorance which you have rightfully pointed out.
As for Apple and ARM it is the operating system that defines the machine. Mac OS is what makes the Mac Book a Mac.
And? The iPad pulled concepts from the iPhone, which pulled concepts from their conceptual tablet which was based on off expertise learned from the iPod and Mac? That DOES NOT make the MacBook a tablet, tablet-like, or tablet-esque. This is simply the natural progression of notebook computers. We can look to the original MacBook Air, which predates the iPad by many years, to see that this shrinking logic board, non-removable battery, single USB port, etc. were enacted. It's silly to say that because we finally have a USB standard that can be used for supplying power to a notebook that it's becoming an iPad. By that logic, Apple's inclusion of revamped frameworks and APIs in iOS that have made Mac OS X more efficient is the Mac becoming an iPhone. That's simply ridiculous.
A non-convertible notebook with a windowed OS, mouse pointer, and no touchscreen is NOT a tablet, with or without a keyboard. An actual tablet computer or convertible, with a keyboard is the "tablet with keyboard" niche.
It's a smaller die on broadwell and low power 5w on the M series. It's not out performing the Haswell range, that will come later this year with skylake. If you want a thermally limited lower power cpu than the previous gen Airs then this is for you.
Excellent. Now it all makes a lot more sense.
Thanks for taking the time to explain.
The downside is incompatibility on the software side with the existing software base compiled to x86 and the loss of Windows compatibility via Bootcamp or parallels.
Also, while Intel has shown that they can play down in the <5W TDP range with ARM, ARM has not yet shown they can scale up to Core i7 performance levels.
That Intel has a 4.5W part that performs almost at Core i5 levels is a significant indicator that the window of opportunity for ARM to displace x86 on the MacBooks closed. A few years back there was the thought that if Intel stumbled on the performance per watt category that ARM would massively displace them in the notebook and server markets. With Core-M the notebook transition seems less likely today and with Xeon-D the same can be said of the server market.
At least for the near future (5-6 years).
1) Apple has done the transitions several times with great success. I don't that today, with an centralized Mac App Store in place, it will be an issue.
2) Why does it need tone Core i7 performance levels? I'd think it would start on the lower-end of the performance scale. Note that the performance of CPUs is not a driving factor for the average customer.
3) I think Core M performance indicates Apple is much more likely to offer an ARM-based traditional "PC".
It would mean laptops (mostly) with tailored made SoC technology.
Look at it this way a SoC is effectively the equivalent of a printed circuit board from the 80's to 90's time frame. It is where engineers now glue their unique designs together. Instead of designing a printed circuit board and a few custome chips to realize most of your platform, you instead design a SoC much like Apple has with the A series chips.
Now why would Apple want to do this for a Laptop? It allows for exactly what they need and want on the SoC instead of settling for Intels generic solutions.
That is all right, it is actually a good question. From the standpoint of people want to see this happen I think it boils down to these issues:
If multicore ARM was the solution, android would have the market cornered. I thought we learned a LONG time ago, that just adding more cores only works with GPUs?
1) Apple has done the transitions several times with great success. I don't that today, with an centralized Mac App Store in place, it will be an issue.
Neither Bootcamp nor Parallels will efficiently run on ARM. Likewise emulating x86 compatibility with a Rosetta like ability was livable for users because the Core Duo outperformed the G5 by the time the transition happened.
In the Ars benchmarks the Core Duo iMac generally outperformed the G5 iMac using universal binaries.
http://arstechnica.com/gadgets/2006/01/imac-coreduo/5/
There was also a very long period where universal binaries did not exist for many apps including many of Apple's own. The existence of the Mac App store does not magically translate into programmer hours required to do even fairly minimal translation, recompile and most importantly retesting.
For the same reason that when Apple transitioned from PowerPC to X86 they did so across the entire computer line as quickly as possible. Software compatibility across the entire line up is hugely important for developers. To leave the MBP, iMac and Mac Pro on x86 while the Macbook is ARM results in the same mess Microsoft ended up with between Windows and Windows RT Surface platforms.
No transition to ARM will occur unless Apple is sure that: a) windows compatibility will be unimportant for business users and b) the top end iMac, MBP and Mac Pro will be faster than the same generation Intel offerings or the heavy professional users will likely jump ship to Windows because Time = Money.
This is a ridiculously silly position. The Core M shows that ARM provides very little advantages over X86 on traditional PC platforms except on price. If Intel was still stuck at 15W+ TDP while ARM could provide similar performance at <5W then there would be a significant case for moving from x86 to ARM.
At 4.5W Intel offers Apple a way forward with fanless designs with long battery life while keeping software compatibility from the thinnest ultra book to the top end Mac Pro.
If anything it indicates that Apple is much more likely to offer an x86-based pro tablet (or convertible) someday running OSX and with a digitizer in addition to multi-touch for creatives. More likely, not likely. Odds are still pretty low but better than zero which is what it was when decent Intel performance was still in the 15W range.
:???: And Apple was foolish to buy PA Semi and think they could ever do anything with ARM compared compared to the incumbents, especially coming out with a 64-bit SoC with a full 64-bit version of iOS years before anyone else. Where is Andrioid's cornering of the AArch64? It doesn't exist because there the benefit isn't there without vertical integration and proper planning. That isn't to say that Apple will offer more cores — who knows, perhaps they'll stay at 2-cores for ever — but I think that such a scenario is unlikely.
What I was addressing was his idea of just taking an iPhone SoC and adding some more cores. I mentioned Android because the trend there seems to be to just add more cores, and in fact, Apple did the opposite starting with the A6 SoC.
Why don't if for example a chip foundry has plans for going from say 30 to 22 to 14 nm, why don't they start designing 14, when the 30 is released? Why is it that they need to go down in steps. I presume there is an actual reason and not just the obvious cynical reason.