What's their to compare? Don'twe know that the fastest Apple processor is slower than the slowest Intel I-processor?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Testing ARM Based Mac Prototypes with Large Magic Trackpad?
- Thread starter MacRumors
- Start date
- Sort by reaction score
What's their to compare? Don'twe know that the fastest Apple processor is slower than the slowest Intel I-processor?
This rumor is based on that Apple presumably has a prototype, which means they are toying with the idea. Do you know what version of an Apple chip this is running, and how it performs. No.
Chrome books are interesting but the OS is a big problem. Great platforms for Linux though. Even the ARM based machines have potential as Linux platforms. There are however real problems with Chrome books that make them a bad buy for many uses. One is the paltry amount of flash included.
By the way I was disappointed with Apple in that the Airs didn't get a flash SSD update because even the Airs are a little tight with storage.
It was painful only for people that wanted it to be painful. People that hot with the program never looked back.
This is an entirely different situation. First we aren't talking the entire product line, at least not anytime soon. Second you make an assumption that the machine would be slower than the Air or iPad. There is a very good possibility that it won't be overall. Third things like Chromebooks, iPads and other devices have clearly demonstrated that i86 support isn't a big deal anymore. This is perhaps the biggest factor here, nobody cares about i86 anymore. It is all about web access, E-Mail and other canned solutions.
So you look at these tasks that people are interested in doing with a computer and you look at how the apps are built and it becomes obvious that an ARM dashed notebook will perform very well for the users Apple would be after with the device. This is especially the case if they can move beyond two cores and the constrained features of the current A7.
In a nut shell i86 doesn't matter to the people Apple would sell these devices to.
----------
Is it? Try running some of those I processors at 1.3 GHZ and see what happens.
Good luck with all your cores on a single-threaded application. Which are still a majority. More cores won't give you faster browser or office performance. We are talking normal usage here, not embarrassingly parallel algorithms.
Taking multi-core performance as the main score, especially on a series of micro benchmarks like Geekbench, is trivialising the discussion. I have access to a supercomputer which contains more then 18000 cores. I bet though, that a MacBook Air would beat it in a browser benchmark...
Mainly because it wastes space and transistors in a laptop portable chip which is not a good thing.
Apple should be able to easily and quickly evolve the A series chips into something that leaves little to be desired at the price point in question. I really don't see the reasons for the concerns everybody has.
That simply isn't true. The vast majority of software on the Mac is threaded, uses processes or in other ways can make use of those cores. Sure you can find examples of demanding apps that aren't threaded but the user picking up a Mac OS based machine will not have a problem with the included apps.
Wrong again. These programs are highly threaded or spawn other processes. It is fair to say that if you went back to a single core on a Mac you would be very disappointed.
While I have to say Geekbench has no place in this discussion, you can't dismiss the importance of cores with a modern operating system and apps. Just about any app accessing the web these days is threaded.
So would an iPad so what is your argument? Apparently you just don't understand how modern operating systems and apps work. Right now 4 cores seem to be the sweet spot for general interactive use. Put somebody with a little skill behind that machine and 4 cores isn't enough. This applies to Intel as well as any other architecture out there. It is extremely easy these days to slow a dual core machine down to a crawl.
----------
Nope it isn't OK because obviously you don't know what is being talked about here. This is a discussion related to a low end laptop, using a high performance ARM CPU. These are not machines designed for legacy software users and would not replace those machines.
That didn't boost the size of intel by several magnitudes. The majority of what Apple produces doesn't use intel chips.
Apples A7 isn't really that far behind.
The problem here is the gap isn't as big as some think. I can see Apple more than doubling performance per core with little effort. Combine that with more cores and you have a very very interesting platform for the low end. People need to remember that the A7 is a cell phone processor with the cores trapped within those limitations. Break free of those limitations and it will hum right along.
----------
As far as development goes the industry has been leaving Intel in herds. Far more development is taking place for ARM these days than i86. Tablets, cell phones, low power laptops, appliances have all gone to ARM. Basically the industry has given Intel the finger and denied it access to the biggest growth markets of the last few years. This is why Intel has suspended desktop development in favor of ultra mobile and has also laid off 5000 people in 2013.
Intel hardware is just a terrible option where engineering is done on the silicon and not the PC board.
Not anytime soon. People in this forum can't seem to grasp the idea of a new class of devices.
Exactly, the bulk of Apples sales go to machines with some of the lowest performance Haswell parts out there.
Here again people seem to forget that A7 only runs at around 1.3GHz in today's machines. Bump that speed into comparable Intel territory and the performance can easily double if the I/O is adjusted to keep up.
Or they could just shut up! Clearly many of them have real i86 needs but more than a few don't have a clue. The vast majority just don't care about i86.
The problem here is the gap isn't as big as some think. I can see Apple more than doubling performance per core with little effort. Combine that with more cores and you have a very very interesting platform for the low end. People need to remember that the A7 is a cell phone processor with the cores trapped within those limitations. Break free of those limitations and it will hum right along.
----------
even if this rumor is 100% true it didn't imply that Apple was leaving Intel behind on all machines.
As far as development goes the industry has been leaving Intel in herds. Far more development is taking place for ARM these days than i86. Tablets, cell phones, low power laptops, appliances have all gone to ARM. Basically the industry has given Intel the finger and denied it access to the biggest growth markets of the last few years. This is why Intel has suspended desktop development in favor of ultra mobile and has also laid off 5000 people in 2013.
Intel hardware is just a terrible option where engineering is done on the silicon and not the PC board.
Okay, for those freaking out thinking Apple is abandoning Intel... consider the fact that one of Apple's key markets for ages has been education. Consider that the one area that they've seen the biggest targeting of product of late that eats into their strength is... education. I work in a school system and notoriously we've been all Mac. That said, with costs always being a concern at the education level... we've begun transitioning to Chromebooks and Google Docs and away from Office and MacOS X.
For Apple to stay in the game, they have to be price-competitive. To do so, is just nigh on impossible with Intel-based machines at that price point. Intel charges way too much per unit in this segment. There's a ton of low-end Chromebooks and many of the stronger competing one's are embracing ARM vs. Atom or Celeron based on a performance per $ metric. Even if you might give up a hair of performance... Apple is not giving up big $ to Intel per unit in having their A7/A8/A# chips produced for this task. They do their own ASICS work, ship it off to Samsung or Taiwan Semiconductor to ramp up production, and benefit heavily in costs per part so they can maintain healthy margins per unit and remain price competitive. Netbook's failed because they didn't do much of anything right. Chromebooks are succeeding because they have enough performance to do the majority of what their particular audience needs. They're not for everybody (yet) but for some audiences they play in, they are an exceptional device value.
For those lamenting the on-board storage... you're missing the point. In most arenas where this device would be beneficial, the ideal is to have little to no on-board storage. Districts like mine have long been using OS X-based laptops with netboot versions of OS X for user accounts and IMHO it's been less than an ideal solution. You have huge hard drives that are used for nothing other than localized system operation (which we generally don't like and it isn't very efficient, you end up with a desktop of shared documents for various students that use it). The net-based booting is abominably slow, inefficient, and just downright unusable at times.
The way the Chromebook operates is actually superior by miles in that it runs a localized system on the laptop itself (off of a smaller flash drive so the machine is pretty snappy) but uses a cloud-based user account (via Gmail, each of our students and staffers has their own account) that links with Google Drive, Google Docs (via Drive) and allows administration management of services and the ability to share (via functionalities clearly cribbed from Google+). It's actually a very ingenious platform for district administration in that we store everything for pennies in the cloud where it's backed up and saved by Google, it's linked to each student's user account, and it enables more uses for each item as we can integrate them into Google (Web) Pages that the kids make and share with their friends, parents and the educators themselves.
The bottomline is... cloud computing at the education level = huge. Huge hard drives, massive amounts of Flash storage, localized storage, etc. = immaterial and actually are outside of the price scopes of what an educational institution is looking for. It costs more money to have all of that, it's excess for their practical needs, it's overkill.
For Apple to remain competitive in education, they have to have a product that competes in education. Having an i3 or i5 or i7 is overkill when administration wants as many devices as it can get on the cheap to put in kids hands, with real keyboards (to teach typing skills) and a way to interact with user services via Cloud Apps. Docs, right now, is eating iWork and Windows' web apps. for dinner right now. Microsoft and Apple's offerings are priced outside of the realms and while they might be good for consumer sales... they miss the boat on mass volume license sales for education. Google gets it.
This move, IMHO, with ARM... is more about Apple trying to compete in this market than it is about trying to compete or replace the current Air, Macbook, or transition the iMac and Mac Pro to ARM. That isn't happening anytime soon. I don't think the performance and storage and the consumer market's mindshare is caught in the cloud services angle as much as IT at the education level is. We're not as concerned strictly on privacy and in fact, the more transparency and oversight that the district has into the student's use of the computers... the better it is from a district administration standpoint. There's still bumps in the road, a desire for many instructors to shut off Google Talk as an example, but overall... the lion's share of what Google has right here is pretty dramatic.
For those that look at Surface with Windows RT as a failure here as a pointer to Apple failing in this realm... it's not the same market by any stretch. Surface RT failed because it was a device without a market. Microsoft built it and shipped it and expected it'd just go out and compete with the iPad. Traditional Windows users were lost about a device that didn't use existing software, that had a different UI (which Windows 8 itself on the desktop has battled as well), and that didn't have alternative versions of existing software bundled in. Had the Surface RT shipped with a free copy of Office on-board sort of like how Apple has bundled iWork and iTunes with new Macs, it might be more of a success. Had there been more marketing embracing cloud services and pushing cloud apps. as replacements, better and more pervasive integration with Microsoft's cloud storage, etc. it might be more successful. That said, it'd also take a lot of consumers giving up on things that they're used to. It'd also require Microsoft to figure out how to make the devices interoperate with their desktop in ways akin to what Apple has groomed it's userbase to be able to do with iTunes on Windows and Mac with regards to the iPhone and iPad.
Not to mention... like I said, if the device is sold in volume to educational institutions and business-related audiences (that aren't wrapped up in personalization, localized file storage, music library and photo library storage, privacy concerns, etc.) and pushes cloud services heavily (iCloud, iWork, iTunes in the Cloud, etc.) with some key revamps (personally, Apple's cloud strategy has been a bust IMHO, I even prefer Play Music to iCloud/iTunes in the Cloud)... it alone is enough market to sustain the device.
I don't look at this as the consumer device to replace the vast majority of users' iMac or MacBook Air or MacBook Pro or MacBook Pro with Retina or whatever. I, instead, look at it as a device that can run localized FatBinary Mac apps. but that is instead Apple's push to roll out more and more iCloud-based products and HTML5-base apps. and is designed more to compete head-on with Chromebooks for education. I can see Apple simultaneously shipping on Intel and ARM for Mac for MANY years to come in the same way that ChromeOS is running on Intel and ARM right now. The difference, of course, is that ChromeOS is nowhere near ready for the heavy lifting that MacOS can do... whereas MacOS can scale down to do everything ChromeOS can do right now (and then some). ARM can do a lot on the low end that Celeron and Atom both do at much higher costs per part. Apple can focus on MacOS ARM shipments for the areas where they need a much more price-competitive unit and push the Intel-equipped MacOS shipments for the audiences where it's more desirable and where client-side apps. and privacy/data storage on a more localized level are not only preferred, but mandatory.
For Apple to stay in the game, they have to be price-competitive. To do so, is just nigh on impossible with Intel-based machines at that price point. Intel charges way too much per unit in this segment. There's a ton of low-end Chromebooks and many of the stronger competing one's are embracing ARM vs. Atom or Celeron based on a performance per $ metric. Even if you might give up a hair of performance... Apple is not giving up big $ to Intel per unit in having their A7/A8/A# chips produced for this task. They do their own ASICS work, ship it off to Samsung or Taiwan Semiconductor to ramp up production, and benefit heavily in costs per part so they can maintain healthy margins per unit and remain price competitive. Netbook's failed because they didn't do much of anything right. Chromebooks are succeeding because they have enough performance to do the majority of what their particular audience needs. They're not for everybody (yet) but for some audiences they play in, they are an exceptional device value.
For those lamenting the on-board storage... you're missing the point. In most arenas where this device would be beneficial, the ideal is to have little to no on-board storage. Districts like mine have long been using OS X-based laptops with netboot versions of OS X for user accounts and IMHO it's been less than an ideal solution. You have huge hard drives that are used for nothing other than localized system operation (which we generally don't like and it isn't very efficient, you end up with a desktop of shared documents for various students that use it). The net-based booting is abominably slow, inefficient, and just downright unusable at times.
The way the Chromebook operates is actually superior by miles in that it runs a localized system on the laptop itself (off of a smaller flash drive so the machine is pretty snappy) but uses a cloud-based user account (via Gmail, each of our students and staffers has their own account) that links with Google Drive, Google Docs (via Drive) and allows administration management of services and the ability to share (via functionalities clearly cribbed from Google+). It's actually a very ingenious platform for district administration in that we store everything for pennies in the cloud where it's backed up and saved by Google, it's linked to each student's user account, and it enables more uses for each item as we can integrate them into Google (Web) Pages that the kids make and share with their friends, parents and the educators themselves.
The bottomline is... cloud computing at the education level = huge. Huge hard drives, massive amounts of Flash storage, localized storage, etc. = immaterial and actually are outside of the price scopes of what an educational institution is looking for. It costs more money to have all of that, it's excess for their practical needs, it's overkill.
For Apple to remain competitive in education, they have to have a product that competes in education. Having an i3 or i5 or i7 is overkill when administration wants as many devices as it can get on the cheap to put in kids hands, with real keyboards (to teach typing skills) and a way to interact with user services via Cloud Apps. Docs, right now, is eating iWork and Windows' web apps. for dinner right now. Microsoft and Apple's offerings are priced outside of the realms and while they might be good for consumer sales... they miss the boat on mass volume license sales for education. Google gets it.
This move, IMHO, with ARM... is more about Apple trying to compete in this market than it is about trying to compete or replace the current Air, Macbook, or transition the iMac and Mac Pro to ARM. That isn't happening anytime soon. I don't think the performance and storage and the consumer market's mindshare is caught in the cloud services angle as much as IT at the education level is. We're not as concerned strictly on privacy and in fact, the more transparency and oversight that the district has into the student's use of the computers... the better it is from a district administration standpoint. There's still bumps in the road, a desire for many instructors to shut off Google Talk as an example, but overall... the lion's share of what Google has right here is pretty dramatic.
For those that look at Surface with Windows RT as a failure here as a pointer to Apple failing in this realm... it's not the same market by any stretch. Surface RT failed because it was a device without a market. Microsoft built it and shipped it and expected it'd just go out and compete with the iPad. Traditional Windows users were lost about a device that didn't use existing software, that had a different UI (which Windows 8 itself on the desktop has battled as well), and that didn't have alternative versions of existing software bundled in. Had the Surface RT shipped with a free copy of Office on-board sort of like how Apple has bundled iWork and iTunes with new Macs, it might be more of a success. Had there been more marketing embracing cloud services and pushing cloud apps. as replacements, better and more pervasive integration with Microsoft's cloud storage, etc. it might be more successful. That said, it'd also take a lot of consumers giving up on things that they're used to. It'd also require Microsoft to figure out how to make the devices interoperate with their desktop in ways akin to what Apple has groomed it's userbase to be able to do with iTunes on Windows and Mac with regards to the iPhone and iPad.
Not to mention... like I said, if the device is sold in volume to educational institutions and business-related audiences (that aren't wrapped up in personalization, localized file storage, music library and photo library storage, privacy concerns, etc.) and pushes cloud services heavily (iCloud, iWork, iTunes in the Cloud, etc.) with some key revamps (personally, Apple's cloud strategy has been a bust IMHO, I even prefer Play Music to iCloud/iTunes in the Cloud)... it alone is enough market to sustain the device.
I don't look at this as the consumer device to replace the vast majority of users' iMac or MacBook Air or MacBook Pro or MacBook Pro with Retina or whatever. I, instead, look at it as a device that can run localized FatBinary Mac apps. but that is instead Apple's push to roll out more and more iCloud-based products and HTML5-base apps. and is designed more to compete head-on with Chromebooks for education. I can see Apple simultaneously shipping on Intel and ARM for Mac for MANY years to come in the same way that ChromeOS is running on Intel and ARM right now. The difference, of course, is that ChromeOS is nowhere near ready for the heavy lifting that MacOS can do... whereas MacOS can scale down to do everything ChromeOS can do right now (and then some). ARM can do a lot on the low end that Celeron and Atom both do at much higher costs per part. Apple can focus on MacOS ARM shipments for the areas where they need a much more price-competitive unit and push the Intel-equipped MacOS shipments for the audiences where it's more desirable and where client-side apps. and privacy/data storage on a more localized level are not only preferred, but mandatory.
More importantly the important apps that people buy these sorts of machines for are threaded or in other ways take advantage of cores.
No to mention those apps do use threads.
It is doubtful of an entry level machine. I do believe that this was the biggest mistake of the reporting. I can't see Apple interested in anything other than a single SoC for this class machine. They could however group CPU complexes into groups of four dirt of like Inteks approach with the new XEONs that use groups of five.
Why not multiple chips starting at four cores. Each upsell model could add another four cores. As to core counts, modern PCs really need a minimal of four cores these days, that is assuming no hyper threading. Given that these cores might be a bit slower more wouldn't hurt but single apps might have a hard time leveraging the extra cores. So if we look at four cores as the sweet spot for most users then additional cores can be sold to more hard core users.
In any event ARM cores are extremely low power it is the caching and memory interface that burns up power. When you leave the constraints of a cell phone additional cores aren't a big deal. Each core is probably less than a watt of burden on the device. So a many core ARM device isn't impossible at all.
The concept doesn't even make sense. Why would apple put 4-8 SOC chips on one device? SOC = system on a chip. That means that a bunch of the features on each chip will be redundant.
I should have asked how much "4-8 64-bit ARM Quad-core processors" (quoted from the article) would cost. I replied to someone who said there would be no price reduction for an ARM-based OS X device.
It also lights a fire under Intel. Do you think Intel is going to be telling Apple about another processor delay if they see Apple with ARM PCs ready to start manufacturing at any time. Apple might intro an ARM laptop at the low end of OS X portables, and let Intel know the same thing could happen to the far more profitable mid and high end products as well if Intel doesn't pick up their pace.
Back in the 10.6 Lion days, most OS X apps were ported to 2 different processors (fat PPC+iA32) with an over 4X performance degradation between the top and bottom Macs in popular use. Most current popular iOS apps are "ported" to 3 or 4 different processor instruction sets, all with very different performances on their respective devices.
The majority of Mac and iOS software customers didn't and still don't know or care.
Many iOS developers build these multi-instructiion-set/fat binaries without even knowing that they are doing it.
Why do so many people think Intel really cares about Apple? Apple Worldwide market share in computers is just 5%. Companies like HP, Dell, Lenovo, Asus etc. buy a lot more processors from Intel than Apple.
What if apple were to use both processors and have an ios like low power mode? They could also provide auto switching like the do with the GPU.
It's more than 5%, and Apple is a big seller of PCs. You might be thinking of Mac vs Windows, but not all Windows computers are made by the same manufacturer; it's split across many. And, also importantly, Apple uses only Intel CPUs, while others seem to use AMD half the time.
----------
Low power mode? It's called Tiger, or at least NOT Mavericks. Or Debian Linux or Ubuntu 3. Before the bloat
Being serious, rather than something unsuitable like iOS, Apple could just have a low power mode that disables animations (I'm looking at you, Exposé), Spotlight, and Time Machine temporarily.
----------
I feel your pain. Especially recently, Apple's been screwing up. Oh well, better than Microsoft any day.
I agree with your post that in order to continue to innovate, they must access the silicon. But my point is, at what cost? It appears (at least initially) that Macs are moving away from being the most flexible machines ever to highly proprietary, locked-down devices once again. I guess it all goes in circles.
I'm sure that this all makes great business sense for Apple. But once again, what will the impact of this transition be on us?
Apple loves to push forward innovation early (sometimes, too early). I think that if they drop the ax and completely eliminate the existing hardware architecture many will jump ship, at least initially.
You have no idea what you are talking about the i7 in the MacBook Pro gets around 15,000 on geekbench not 7,000. You are way off base.
----------
You are comparing the A7 to 2012 MacBook Pro that was release 2 years ago. You think they can put 4 to 8 Quad Core A7 into a laptop, it isn't that simple, and even then it would be multiple generations behind in speed.
He is probably talking about the Dual Core 2.8Ghz Core i7 in the 13" MBPr which is around 7000 in Geekbench 3 (64bit).
Actually the A7 in the iPhone 5S scores 2400.. not quite double the 1300 of the A6 or 1400 of the A6X. Geekbench is eh. A lot of the A7's performance boost in Geekbench actually comes from cryptography improvements (e.g. AES), which don't make much difference in 99% of applications. And the graphics wouldn't be even close.
Maybe. But not if it's supposed to be so ultra-thin that it's even slimmer than the current MBA. By the way, it's only a matter of time before quad core is available in a 13" form factor as well. And your mention of doubling the clockspeed would MORE than double the power consumption of the chips. Way more than double it. It might even quadruple it, and then if you double the number of cores, well, who's to say that it wouldn't exceed the 15W package that Intel offers for its low-power chips in the MBA?
It's not as simple as you seem to think it is.
Maybe.
Apple's shift from PowerPC to Intel made sense because so much of the world already ran on Intel. That's not true for ARM outside of mobile space.
It does not matter whether its threaded or not. The question is how the performance-critical workflow is distributed over multiple cores. All the user interface code, where the JavaScript is executed, DOM manipulated, website redrawn etc. — its all happening on one thread. Same for office applications. What good would it be if I have 16 slower cores if the application I am using right performs most of the important work in a single thread?
I feel that this discussion is getting a bit off-topic, so I would like to remained you where it started — with the discussion of the ARM performance and whether it would be sufficient for desktop needs. My point was that having multiple cores is not a panacea if the individual cores do not deliver enough single-core performance. Take away the turbo boost features of modern CPUs and you will be amazed how much the performance goes down. This is because even with modern OS, the workloads are not distributed evenly over the cores — usually, there is always one that does most of the work.
Again, I am not trying to downplay the importance of the multiple cores! I am simply pointing out that the single-core performance is still absolutely crucial. A multi-core CPU won't bring you much when you have Office and a browser open at the same time — as it spends majority of its time in idle anyway. A multi-core CPU will help you if you are working in the office and encoding a video file in the background. There is a crucial difference there.
Just some food for thought: all other things (e.g. architecture) equal, I would always take a single-core 4Ghz CPU a quad-core 0.5Ghz CPU. The reason — with the single-core one I can have max performance whenever I need it — both in single-threded and multi-threaded tasks. With the multi-core one, I can have the maximal CPU utilisation only with heavily parallel workflows which are evenly distributed over the cores.
Last edited:
This is not going to be about the CPU. Yes, intel will be faster for some time. But they are not giving apple what they want.
Apple wants a fast cpu, with an amazing GPU, low power consumption, for a good price. Out of the 4, intel can deliver only 1.
Apple will probably want to do a retina macbook in a fanless design, with great battery life and competitive price. Intel chips don't allow for this design(afaik).
Apple wants a fast cpu, with an amazing GPU, low power consumption, for a good price. Out of the 4, intel can deliver only 1.
Apple will probably want to do a retina macbook in a fanless design, with great battery life and competitive price. Intel chips don't allow for this design(afaik).