Let's look at the facts. Intel's x86-64 CPUs run significantly hotter than ARM processors. True, ARM processors trail Intel's x86-64 ULV processors (used by MBA) by a great deal. Specifically, iPad 2 is about 3 times slower than entry level MBA. Although I wouldn't go far and say the performance gap is closing, it is certainly narrowing.
Meanwhile, at 18W of TDP, Intel's Sandy Bridge ULV processors (which are destined for MBA) is even hotter than Core 2 ULV used on MBA (10W). iPad 2 has mere 0.5W TDP.
At some point, Apple's ARM processors will be fast enough for mainstream notebooks, starting with MBA. It would be foolish of Apple to not invest in ARM-flavor Mac OS X, just as it invested in both PowerPC and Intel Mac OS X nearly a decade ago.
Let's look at the facts. Intel's x86-64 CPUs run significantly hotter than ARM processors. True, ARM processors trail Intel's x86-64 ULV processors (used by MBA) by a great deal. Specifically, iPad 2 is about 3 times slower than entry level MBA. Although I wouldn't go far and say the performance gap is closing, it is certainly narrowing.
Meanwhile, at 18W of TDP, Intel's Sandy Bridge ULV processors (which are destined for MBA) is even hotter than Core 2 ULV used on MBA (10W). iPad 2 has mere 0.5W TDP.
At some point, Apple's ARM processors will be fast enough for mainstream notebooks, starting with MBA. It would be foolish of Apple to not invest in ARM-flavor Mac OS X, just as it invested in both PowerPC and Intel Mac OS X nearly a decade ago.
Apple could seriously under clock the i3 and match ARM's power envelope and still blow it away on performance. And when Ivy Bridge comes out, Intel's power consumption drops by 50%.
If Apple wanted a slow laptop that didn't use much energy, they could get there without having to switch at ARM.
Perhaps as Apple do with dual, optional graphics systems now in MBPs, there is a possibility Apple plan to 'include' an ARM as well as whatever main CPU is in use. This might allow iOS to run at full speed in a window along with OS X.
Now that would be more sane. A quick boot/uber low power consumption mode running iOS for quick tasks, and OSX on a beefier processor for heavier tasks.
Interesting read -- but they gave no reasons to support their claim.
It may be that article and this thread are based on the concepts:
1) much laptop and some desktop usage will evolve to be what is being done on the iPad -- not use heavy lifting apps
2) use of something like the Chrome OS combined with cloud apps -- to approach/approximate heavy lifting
3) rewrite a few, key, high-use apps (Word, Excel, etc.) for ARM to satisfy all needs for many users.
4) write (or rewrite) specialized apps to exploit the ARM acrhitecture for specific industries.
As to the latter -- it appears that some existing iPad (ARM) apps already satisfy many of the needs of medical, education, etc.
Here's a hypothetical case:
Apple is rewriting Final Cut Studio from the ground up (non-standard UI, Carbon code).
What if the FCPX replacement to FCS can be made to run as well * on current ARM as on current x86?
* AFAIK, there is nothing in FCS that requires x86 power -- rather it needs GPU, Cores/Threads and h264 encoding capability.
Then it is possible that the needs of an industry segment (Video Post) can [largely] be satisfied by ARM based tablets, laptops and desktops.
It is interesting that a pretty good iMovie is already available for the iPad.
It's not too far fetched to think that the next Mac iMovie will run on both x86 and ARM.
So what you might ask -- Apple can build some ARM laptops and Desktops that can run iMovie and Final Cut (as well as or better than x86)?
What if that could satisfy an industry segment?
You laugh...
A friend who is an executive for one of the broadcast networks -- told me this (emphasis mine):
Quote:
FCP is actually a powerful program but I still think iMovie does the job for most less than 10 minutes productions. Its [FCP] real value is revealed when using After Effects or Motion, integrated tight in the production. Content is still king and video 'direction' makes a video look pro... not really the 'editing' tools in most cases.
If you ever travel to NY I would love to give you a tour of some of the edit suites and see how the product is integrated in the workflow. FCP is not the main edit tool, however as a FCP fan you will see its value when connected to graphic virtual sets and tapeless video ingest servers. Pretty amazing in capable hands. But you will also see how simple on & off-line systems (equiv. to iMovie) does the bulk of the work.
[ ... ] Even Microsoft has plans for the ARM architecture in the future, as mobile devices offer longer battery life with the low-power chips. The Redmond, Wash., software giant revealed at this year's Consumer Electronics Show that the next version of its desktop operating system, Windows 8, will run on the ARM's architecture. [ ... ]
Let's see now. Microsoft ported Windows NT to several RISC platforms. But failed to bring Windows to the mass market on RISC. Why? Because Office, their other cash cow, had pervasive data alignment issues thanks to being developed on the weird, hackish x86 legacy architecture: http://en.wikipedia.org/wiki/Windows_NT
On the other hand, Mac OS X was (and its apps were) written on RISC and later ported to CISC. Yes, it took years, but Apple has years until ARM chips are ready to run Mac OS X. Or whatever its successor will be called.
If any company can plan ahead, it's Apple. I'm sure the Mac OS X-on-ARM scenario was mapped out even before the PA Semi acquisition. There can be no doubt that Apple's Holy Grail is a unified OS for Macs and iDevices. (No, not multi-touch iMacs. That will never happen.) A common core OS that can drive UIs for display sizes from iPod to HDTV and a single unified App Store would give Apple a huge advantage.
(You'll note that I said "Holy Grail." I didn't say "future plans." Just wanted to deflect any knee-jerk reactionary "keep the status quo" comments.)
AIR, the top PASC talent left Apple when they didn't get big stock options.
I suspect that, for practical reasons, Intel will have to participate in creating an architecture that will subsume the x86.
Probably, but please stop trying to bring practicality into my dream world. Intel's participation in
design wouldn't be my first choice, because then the new architecture would be something that Intel would sell to everyone. I would prefer that the new architecture be developed by Apple, so that their optimization of hardware with OS could begin at the architectural level, giving them a huge advantage over users of commodity processors with other operating systems. If Apple wanted to use Intel as a dumb foundry, that would be fine.
Apple could seriously under clock the i3 and match ARM's power envelope and still blow it away on performance. And when Ivy Bridge comes out, Intel's power consumption drops by 50%.
If Apple wanted a slow laptop that didn't use much energy, they could get there without having to switch at ARM.
Um, no. Because even at idle, Sandy Bridge ULV processors consume more power than iPad's A5 running at maximum.
Each iteration of ARM is quickly approaching the level where it would be entirely suitable for general mainstream notebooks. ARM Cortex-A9 isn't there obviously, but 64-bit version of whatever succeeds ARM Cortex-A15 could be entirely suitable for MacBook Air.
I am not saying ARM will be the only CPU architecture for Mac OS X. But Mac OS X had supported dual binaries since 10.4 (when it added Universal Binary). 10.6 drops support for PowerPC and 10.7 will drop support for 32-bit Intel (x86). It won't be as major challenge to introduce another target as you think.
Let's see now. Microsoft ported Windows NT to several RISC platforms. But failed to bring Windows to the mass market on RISC. Why? Because Office, their other cash cow, had pervasive data alignment issues thanks to being developed on the weird, hackish x86 legacy architecture: http://en.wikipedia.org/wiki/Windows_NT
On the other hand, Mac OS X was (and its apps were) written on RISC and later ported to CISC. Yes, it took years, but Apple has years until ARM chips are ready to run Mac OS X. Or whatever its successor will be called.
If any company can plan ahead, it's Apple. I'm sure the Mac OS X-on-ARM scenario was mapped out even before the PA Semi acquisition. There can be no doubt that Apple's Holy Grail is a unified OS for Macs and iDevices. (No, not multi-touch iMacs. That will never happen.) A common core OS that can drive UIs for display sizes from iPod to HDTV and a single unified App Store would give Apple a huge advantage.
(You'll note that I said "Holy Grail." I didn't say "future plans." Just wanted to deflect any knee-jerk reactionary "keep the status quo" comments.)
I believe that the NextStep OS Apple got when it bought NeXT was already running on x86.
They ported it to PPC and released it as OS X.
I believe that Apple did parallel development, behind the scenes, to the x86 version to keep it consistent with the PPC version.
I suspect that Apple would do the same with an ARM version of Mac OS X -- if they chose to implement it.
I believe that the NextStep OS Apple got when it bought NeXT was already running on x86.
They ported it to PPC and released it as OS X.
I believe that Apple did parallel development, behind the scenes, to the x86 version to keep it consistent with the PPC version.
I suspect that Apple would do the same with an ARM version of Mac OS X -- if they chose to implement it.
Some good posts, Dick.
I noted the earlier post when someone noted Arm starting life out on the Acorn computers. Yes. Desktop computers. It pretty much kicked the snot out of Amiga and PC computers back then. It was very, very fast.
The stuff in the iPad 2. A 'second gen' chip...is what? 'Four' times faster? Twice as fast? Take your pick. It's significant. Try to imagine that evolution to 2013. The gpu power is 9 times faster now? Imagine PVR in 2013?
It's pretty academic saying what Intel is now or will be then. It's more a case of where Apple is heading with the iOS platform and frankly, the Mac OS platform is going to blow bubbles keeping pace with the growth of the iOS platform. And we clearly haven't seen the next few cards of Apple's iOS hand played yet. A tv? A bigger tablet? An expanded range of sizes... An iOS 'Air' or 'iMac'. Who knows.
An iPad can already do nine tenths of what I use my computer for...and do it very, very quickly. Having played with an iPad 2...it's very quick. Imagine this with a retina display in 2012? With yet another leap in cpu and gpu performance? Gulp.
Double it's size and you have a very workable iOS 'mini iMac' that takes iOS and Apple into the unchartered territory.
You only have to look at the video effects on the latest iPad 2. It's pretty astonishing in something so thin. The gpu performance is impressive. It's closing in on the PS3/xbox 360.
You could pretty much stick a keyboard on it now. Oh. You can dock with with a keyboard.
With the 'concurrent' WWDC in June featuring Mac Os X and iOS development in tandem...
You'd have to be in serious denial to see this is classic Apple at work here.
I can easily see a super thin iMac tablet equivalent. iOS apps dwarf those of the Mac OS. (Same thing but with the 'fat' cut off?)
Just look at how many iPod, iPhone, iPads have been sold. Getting on for 200 million? More? It eclipsed Mac total sales and installed base some time ago.
Never say 'never' with Apple. I remember when the Intel news dropped. They'll flog the Mac for all it's worth. But I think the ongoing amalgamation is happening before our eyes.
It will be very interesting to see where we end up in two years time.
Comments
Meanwhile, at 18W of TDP, Intel's Sandy Bridge ULV processors (which are destined for MBA) is even hotter than Core 2 ULV used on MBA (10W). iPad 2 has mere 0.5W TDP.
At some point, Apple's ARM processors will be fast enough for mainstream notebooks, starting with MBA. It would be foolish of Apple to not invest in ARM-flavor Mac OS X, just as it invested in both PowerPC and Intel Mac OS X nearly a decade ago.
Let's look at the facts. Intel's x86-64 CPUs run significantly hotter than ARM processors. True, ARM processors trail Intel's x86-64 ULV processors (used by MBA) by a great deal. Specifically, iPad 2 is about 3 times slower than entry level MBA. Although I wouldn't go far and say the performance gap is closing, it is certainly narrowing.
Meanwhile, at 18W of TDP, Intel's Sandy Bridge ULV processors (which are destined for MBA) is even hotter than Core 2 ULV used on MBA (10W). iPad 2 has mere 0.5W TDP.
At some point, Apple's ARM processors will be fast enough for mainstream notebooks, starting with MBA. It would be foolish of Apple to not invest in ARM-flavor Mac OS X, just as it invested in both PowerPC and Intel Mac OS X nearly a decade ago.
Apple could seriously under clock the i3 and match ARM's power envelope and still blow it away on performance. And when Ivy Bridge comes out, Intel's power consumption drops by 50%.
If Apple wanted a slow laptop that didn't use much energy, they could get there without having to switch at ARM.
How about a laptop with an ARM AND x86 CPU.
Great minds ...
How about a laptop with an ARM AND x86 CPU.
Now that would be more sane. A quick boot/uber low power consumption mode running iOS for quick tasks, and OSX on a beefier processor for heavier tasks.
ARM chips to own 13% of PC market by 2015: IDC
http://www.marketwatch.com/story/arm...idc-2011-05-05
Interesting read -- but they gave no reasons to support their claim.
It may be that article and this thread are based on the concepts:
1) much laptop and some desktop usage will evolve to be what is being done on the iPad -- not use heavy lifting apps
2) use of something like the Chrome OS combined with cloud apps -- to approach/approximate heavy lifting
3) rewrite a few, key, high-use apps (Word, Excel, etc.) for ARM to satisfy all needs for many users.
4) write (or rewrite) specialized apps to exploit the ARM acrhitecture for specific industries.
As to the latter -- it appears that some existing iPad (ARM) apps already satisfy many of the needs of medical, education, etc.
Here's a hypothetical case:
Apple is rewriting Final Cut Studio from the ground up (non-standard UI, Carbon code).
What if the FCPX replacement to FCS can be made to run as well * on current ARM as on current x86?
* AFAIK, there is nothing in FCS that requires x86 power -- rather it needs GPU, Cores/Threads and h264 encoding capability.
Then it is possible that the needs of an industry segment (Video Post) can [largely] be satisfied by ARM based tablets, laptops and desktops.
It is interesting that a pretty good iMovie is already available for the iPad.
It's not too far fetched to think that the next Mac iMovie will run on both x86 and ARM.
So what you might ask -- Apple can build some ARM laptops and Desktops that can run iMovie and Final Cut (as well as or better than x86)?
What if that could satisfy an industry segment?
You laugh...
A friend who is an executive for one of the broadcast networks -- told me this (emphasis mine):
FCP is actually a powerful program but I still think iMovie does the job for most less than 10 minutes productions. Its [FCP] real value is revealed when using After Effects or Motion, integrated tight in the production. Content is still king and video 'direction' makes a video look pro... not really the 'editing' tools in most cases.
If you ever travel to NY I would love to give you a tour of some of the edit suites and see how the product is integrated in the workflow. FCP is not the main edit tool, however as a FCP fan you will see its value when connected to graphic virtual sets and tapeless video ingest servers. Pretty amazing in capable hands. But you will also see how simple on & off-line systems (equiv. to iMovie) does the bulk of the work.
It's a conflict of interest for them. By making chips for Apple, Intel is indirectly killing its own market share in the PC business.
[ ... ] Even Microsoft has plans for the ARM architecture in the future, as mobile devices offer longer battery life with the low-power chips. The Redmond, Wash., software giant revealed at this year's Consumer Electronics Show that the next version of its desktop operating system, Windows 8, will run on the ARM's architecture. [ ... ]
Let's see now. Microsoft ported Windows NT to several RISC platforms. But failed to bring Windows to the mass market on RISC. Why? Because Office, their other cash cow, had pervasive data alignment issues thanks to being developed on the weird, hackish x86 legacy architecture: http://en.wikipedia.org/wiki/Windows_NT
On the other hand, Mac OS X was (and its apps were) written on RISC and later ported to CISC. Yes, it took years, but Apple has years until ARM chips are ready to run Mac OS X. Or whatever its successor will be called.
If any company can plan ahead, it's Apple. I'm sure the Mac OS X-on-ARM scenario was mapped out even before the PA Semi acquisition. There can be no doubt that Apple's Holy Grail is a unified OS for Macs and iDevices. (No, not multi-touch iMacs. That will never happen.) A common core OS that can drive UIs for display sizes from iPod to HDTV and a single unified App Store would give Apple a huge advantage.
(You'll note that I said "Holy Grail." I didn't say "future plans." Just wanted to deflect any knee-jerk reactionary "keep the status quo" comments.)
This is why Intel wont supply Apple with chips.
It's a conflict of interest for them. By making chips for Apple, Intel is indirectly killing its own market share in the PC business.
So if Apple were to quit making Macs right this minute, you think Intel would sell more chips?
No, I think Intel is killing their 2020+ business by sticking with x86.
AIR, the top PASC talent left Apple when they didn't get big stock options.
I suspect that, for practical reasons, Intel will have to participate in creating an architecture that will subsume the x86.
Probably, but please stop trying to bring practicality into my dream world. Intel's participation in
design wouldn't be my first choice, because then the new architecture would be something that Intel would sell to everyone. I would prefer that the new architecture be developed by Apple, so that their optimization of hardware with OS could begin at the architectural level, giving them a huge advantage over users of commodity processors with other operating systems. If Apple wanted to use Intel as a dumb foundry, that would be fine.
Apple could seriously under clock the i3 and match ARM's power envelope and still blow it away on performance. And when Ivy Bridge comes out, Intel's power consumption drops by 50%.
If Apple wanted a slow laptop that didn't use much energy, they could get there without having to switch at ARM.
Um, no. Because even at idle, Sandy Bridge ULV processors consume more power than iPad's A5 running at maximum.
Each iteration of ARM is quickly approaching the level where it would be entirely suitable for general mainstream notebooks. ARM Cortex-A9 isn't there obviously, but 64-bit version of whatever succeeds ARM Cortex-A15 could be entirely suitable for MacBook Air.
I am not saying ARM will be the only CPU architecture for Mac OS X. But Mac OS X had supported dual binaries since 10.4 (when it added Universal Binary). 10.6 drops support for PowerPC and 10.7 will drop support for 32-bit Intel (x86). It won't be as major challenge to introduce another target as you think.
Let's see now. Microsoft ported Windows NT to several RISC platforms. But failed to bring Windows to the mass market on RISC. Why? Because Office, their other cash cow, had pervasive data alignment issues thanks to being developed on the weird, hackish x86 legacy architecture: http://en.wikipedia.org/wiki/Windows_NT
On the other hand, Mac OS X was (and its apps were) written on RISC and later ported to CISC. Yes, it took years, but Apple has years until ARM chips are ready to run Mac OS X. Or whatever its successor will be called.
If any company can plan ahead, it's Apple. I'm sure the Mac OS X-on-ARM scenario was mapped out even before the PA Semi acquisition. There can be no doubt that Apple's Holy Grail is a unified OS for Macs and iDevices. (No, not multi-touch iMacs. That will never happen.) A common core OS that can drive UIs for display sizes from iPod to HDTV and a single unified App Store would give Apple a huge advantage.
(You'll note that I said "Holy Grail." I didn't say "future plans." Just wanted to deflect any knee-jerk reactionary "keep the status quo" comments.)
I believe that the NextStep OS Apple got when it bought NeXT was already running on x86.
They ported it to PPC and released it as OS X.
I believe that Apple did parallel development, behind the scenes, to the x86 version to keep it consistent with the PPC version.
I suspect that Apple would do the same with an ARM version of Mac OS X -- if they chose to implement it.
I believe that the NeXTSTEP OS was already running on x86 when Apple paid NeXT to take over their brand name.
Fixed that for you.
Could Intel apply it's new 22nm TriGate process to ARM chips? That would be sweet.
AMD's had this trigate technology for several years and will be introducing it before Intel.
Fix that for you.
Both NEXTSTEP and NeXTStep are correct.
Both NEXTSTEP and NeXTStep are correct.
That isn?t what I fixed.
All this talk of running Mac OS X on ARM...
Through the years, one of the biggest issues with OS X has been:
FFFinder!
In the iOS implementation of OS X, Apple resolved this in typical Apple fashion -- they got rid of it!
Ha!
...Now, if iOS 5 can only find a way to share/access files among apps.
If they can do this elegantly...
They could port the solution back to Mac OS X and FFFinder!
I believe that the NextStep OS Apple got when it bought NeXT was already running on x86.
They ported it to PPC and released it as OS X.
I believe that Apple did parallel development, behind the scenes, to the x86 version to keep it consistent with the PPC version.
I suspect that Apple would do the same with an ARM version of Mac OS X -- if they chose to implement it.
Some good posts, Dick.
I noted the earlier post when someone noted Arm starting life out on the Acorn computers. Yes. Desktop computers. It pretty much kicked the snot out of Amiga and PC computers back then. It was very, very fast.
The stuff in the iPad 2. A 'second gen' chip...is what? 'Four' times faster? Twice as fast? Take your pick. It's significant. Try to imagine that evolution to 2013. The gpu power is 9 times faster now? Imagine PVR in 2013?
It's pretty academic saying what Intel is now or will be then. It's more a case of where Apple is heading with the iOS platform and frankly, the Mac OS platform is going to blow bubbles keeping pace with the growth of the iOS platform. And we clearly haven't seen the next few cards of Apple's iOS hand played yet. A tv? A bigger tablet? An expanded range of sizes... An iOS 'Air' or 'iMac'. Who knows.
An iPad can already do nine tenths of what I use my computer for...and do it very, very quickly. Having played with an iPad 2...it's very quick. Imagine this with a retina display in 2012? With yet another leap in cpu and gpu performance? Gulp.
Double it's size and you have a very workable iOS 'mini iMac' that takes iOS and Apple into the unchartered territory.
You only have to look at the video effects on the latest iPad 2. It's pretty astonishing in something so thin. The gpu performance is impressive. It's closing in on the PS3/xbox 360.
You could pretty much stick a keyboard on it now. Oh. You can dock with with a keyboard.
With the 'concurrent' WWDC in June featuring Mac Os X and iOS development in tandem...
You'd have to be in serious denial to see this is classic Apple at work here.
I can easily see a super thin iMac tablet equivalent. iOS apps dwarf those of the Mac OS. (Same thing but with the 'fat' cut off?)
Just look at how many iPod, iPhone, iPads have been sold. Getting on for 200 million? More? It eclipsed Mac total sales and installed base some time ago.
Never say 'never' with Apple. I remember when the Intel news dropped. They'll flog the Mac for all it's worth. But I think the ongoing amalgamation is happening before our eyes.
It will be very interesting to see where we end up in two years time.
The writing is on the wall.
Lemon Bon Bon.
Originally Posted by Dick Applebaum
I believe that the NeXTSTEP OS was already running on x86 when Apple paid NeXT to take over their brand name.
Fixed that for you.
Both NEXTSTEP and NeXTStep are correct.
That isn?t what I fixed.
HaHaHa!