Apple Testing ARM Based Mac Prototypes with Large Magic Trackpad?

[diddl14]

Good luck convincing (again) all the s/w houses to write their s/w for a different processor. It would be a suicide. Luckily, I believe this won't happen.

Xcode would simply spit out a single binary supporting both intel and arm cup's. That part is a meanwhile a no-brainer and requires no effort at all from the developers.

Key would be real fast x86 simulator for legacy software and Virtual Machines. The A8, combined with Apples LLVM JIT technology might actually be able to deliver exactly that...

If they nail that one down, the migration could be seamless for the bulk of Mac users. Only exception I can think of is bootcamp. Anything else, including kernel extensions could be translated on the fly from x86 to arm code using llvm.

 

[Woyzeck]

Please, not again.

Apple has been strongest when they opened their platform and became more compatible with the rest of the world and they shrinked when they raised their fence too high.

Now we have non-working Samba, more-or-less unusable Exchange integration and perhaps a total niche processor in the future ? C'mon, Apple.

 

[Serban]

MacBidouille have always been right.
They have real inside sources. They posted PowerMac G5 specs over a year before it was released for example.

ARM based Mac prototypes have been around for at least 3 years.

People need to look at benchmarks: A7 is faster per mhz than Intel i7ULV.
But that is not the most important.

By controlling the CPU like Apple does with iOS Apple can optimize the CPU.
For example A5 30% of the die area is for specific Apple stuff like DSPs for Siri.
This would give Apple a huge competitive advantage against generic X86 that relies on brute force instead of optimization.

And the price:
A7 costs Apple 18 dollar to produce. X86ULV + motherboard cost 350-400 dollars. Intel is abusing it monopoly in X86. Thats why X86 have stalled since 2006 when AMD could not compete.

2006-2014: Intel single threaded performance is up 50%.
50% in 8 years!

ARM/Apple Aseries is up over 4000% in performance since 2007.

ARM have at least 10 companies that create custom ARM cores and compete. Since ARM just charge a licensing fee = its cheap. We get competition that don’t exist in X86, we get REAL 64bit instead of X86-64bit extensions.

(its amazing that X64 is 3% slower than 32bit in Windows. Real 64bit gains 20-30% performance by recompile stuff in 64bit. Real computes moved to 64bit in 1990. Not because of the myth 4gig memory, but because of performance)

Intel charges 4400 dollar for its fastest CPU. It cost Intel less than 300 dollar to produce that CPU. (People who for example think Nvidia/AMD high end GPUs are expensive: They cost more to produce than 10 core/20 thread Xeon. Largest Xeon is 470-480mm2. Nvidia/AMD are over 500mm2. Intel have no competition = charge 4400dollars. Nvidia/AMD compete and sell its GPUs without real profit. The profit for Nvidia/AMD is their compute/WS versions of the same card.)

Please die X86. 40 years old. Never good...

yes man but what is happeninig for those americans who play WOW/Starcraft etc on their macs...there are apps only Intel/Amd. And lets not forgot that are still apps on MAC that dont worl and you need windows

 

[na1577]

I'm curious what Apple's rationale for this would be. Windows RT is dying a slow death because Windows 8.1 proper performs well enough on Atom-based tablets. Because of that, there's no good reason for RT to exist. Intel has made massive gains in efficiency in recent years which is why we now have MacBook Air's with "all-day" battery life. We've reached a point where efficiency is adequate for most uses so there isn't a dire need for a more lightweight platform like ARM.

The only thing I can think of is that Apple is trying to create a low cost notebook. We've seen the MacBook Air drop below the $1K mark recently, but what if Apple had a $400-$600 ARM-based Mac? That would fly off the shelves, especially in emerging markets. I could see it running a hyrbid iOS/OS X operating system. It would be more simple, like iOS, but with added bells and whistles (and no touch) to behave more like a regular computer. It would eat up any momentum Chromebooks have amassed in the past few years. A lot of people don't need everything OS X has to offer, and that's why the iPad has been so successful in the post-PC era. This might be the best way for Apple to compete in the entry level market that they have generally avoided being a part of.

Apple is still in a position where it can make major moves and the industry will follow. They don't have to worry about a device like this being unable to attract developers like Windows RT; if there's an Apple platform, developers will write apps for it and the customers will come along with them. I think Apple has a really fantastic opportunity here and I'd like to see them do something about it.

 

[Cloudsurfer]

This will eventually happen. Apple didn't start making chhips just for iOS devices, we will see a switch to ARM within the next 10 years. And to be honest I don't mind. The battery life on an ARM MacBook Pro would be amazing.

 

[IscariotJ]

yes man but what is happeninig for those americans who play WOW/Starcraft etc on their macs...there are apps only Intel/Amd. And lets not forgot that are still apps on MAC that dont worl and you need windows

If this thing does see the light of day, I'd put money on it being a new product line, or an extension of the iPad line. Surely you're not thinking that it'll replace the Intel Macs ( though I bet Apple have run iOS revenue vs Mac revenue )?

I'm sure the grand plan is for Apple to remove it's reliance on as many third party suppliers as it can, and use it's own designs / components, but the culmination of that is years away....

 

[MacHiavelli]

A lot of the time people claim Intel-based computers are needed because users need Windows and Office.

This claim is called into question by three facts...

1: Most people don't actually use Office, and certainly need something more efficient and professional than its bloated and clunky interface. http://appleinsider.com/articles/14/05/03/most-office-workers-arent-actually-using-microsoft-office

2: If *some* people need Office, they can already get it via ARM-based apps by MS and others. Those people don't need Intel-based PCs.

3: We are all increasingly using web-based apps and cloud services, which are processor independent.

Given the inefficnency of Windows machines and the high costs of tech support and software licensing, why should companies carry on wasting money on MS products when Apple solutions are the cheaper and more efficient option?

Certainly, some companies are very slow to change and/or don't have the vision to change from MS to Apple products. But that's not Apple's problem. If Apple switches to ARM processors, the vast majority of users won't care...the benefit of universal apps will be huge.

The future certainly isn't Windows or Office. Apple just has to decide if ARM or Intel will be best for future machines.

 

[4God]

...The only thing I can think of is that Apple is trying to create a low cost notebook. We've seen the MacBook Air drop below the $1K mark recently, but what if Apple had a $400-$600 ARM-based Mac? That would fly off the shelves, especially in emerging markets. I could see it running a hyrbid iOS/OS X operating system. It would be more simple, like iOS, but with added bells and whistles (and no touch) to behave more like a regular computer. It would eat up any momentum Chromebooks have amassed in the past few years. A lot of people don't need everything OS X has to offer, and that's why the iPad has been so successful in the post-PC era. This might be the best way for Apple to compete in the entry level market that they have generally avoided being a part of.

Hence the cheaper iMac rumors...

 

[MacHiavelli]

Delete... Duplication.

 

[JoEw]

The idea of having 4-8 seperate processors is a neat idea, in theory, they could control how much power is needed and then only use a more precise amount as demanded by the operating system. People also forget switching to ARM would allow all ipad apps to run on this mac, and then the App Store could make it's way to mac, and developers who already have an ipad/iphone app just need to code a mac version of the binary optimized for mouse/trackpad input.

This found actually make MORE software available to mac platform all are lower price and all those developers of iphone and ipad apps could pretty easily leverage the platform.

I need some stock

 

[diddl14]

Can anyone confirm this? I always hear it would be way harder than that. Wasn't PPC to Intel way more of a mess?

PPC to Intel was Apple's first time migrating to another architecture and way before iOS/ARM came along. x86 was new to the Apple eco system so clearly had a bit of a learning curve.

Both Intel and ARM code are nowadays highly mature. More important, Apple has all the developer tools in place to simply support multiple architectures in a single binary. Unlike with the transition from PPC to X86 or form 32 to 64 bit, there is typicallly nothing that needs to be done at code level. It really is just recompile and go.

Only exception might be some exotic optimisation using X86 assembly language. Apps that took the effort to optimise at that level, likely also have a the knowhow and resources to do the same for ARM (if they have not already done some when porting their app to iPad).

And for apps that aren't recompiled, on-the-fly llvm translation might do the job. Effectively that would enable all existing apps to run without the developers needing to do anything at all.

 

[MacHiavelli]

Delete ... Duplication.

 

[IscariotJ]

PPC to Intel was Apple's first time migrating to another architecture and way before iOS/ARM came along. x86 was new to the Apple eco system so clearly had a bit of a learning curve.

Actually, it wasn't. They switched from Motorola 68k to PPC....

 

[barkmonster]

What if they manage to emulate X86 code on an ARM chip that's on the level of the performance of the current 1.4Ghz i5s but offers better native performance of ARM-optimsed OS apps like the Finder etc... and it's a hybrid iOS/X86 system that replaces the Macbook Air?

Not a complete platform CPU shift, just a compromise at the entry level that acts as a springboard between iOS and Mac OS X for people who can't justify the expense of any current Mac and want more than a Chromebook or Netbook?

If Apple shift the entire Mac range to ARM, they're a fool because that's a massive step backwards regardless of ARM biting at the lower-lower end of Intels CPU offerings. There's no ARM XEON killer, no ARM Quad or Hex i7 killer and add a Rosetta style emulation layer into the mix for application compatibility and it would literally kill the viablilty of using a Mac in the first place for existing Mac owners and Windows switchers alike.

Don't bother "educating me" about universal binaries either. I know what they are and I know ALL current Mac software will be running under emulation for a long time before native ARM applications are about. It's taken over half a decade for 64bit applications to become more common and systems have offered the ability to upgrade past 4Gb since the G5!

 

[opeter]

If that's true ... well, the x86/x64 will be emulated then, or what is Apple thinking?

The only reason, many people did buy Apple's PCs in the past (aka.: since 2005) were that they can run Windows operating system (x86/x64) and application written for it on the computers.

If this will change, well ... this can be a real drawback.

Anyway, let's wait and see.

 

[xmichaelp]

I pray for the day the lawsuit is filed to prevent Apple from bundling an OS with Safari. I stopped using Safari when the OS was in the early days of 10. Tiger I think, Safari had so many issues, even today, I refuse to use it.

All your posts in this thread are just nonsensical babble. Seriously, what are you talking about? How is this relevant? And safari is the best browser on OS X easily.

 

[JoEw]

Apple dominates business in tablet/phones why don't they extend the same ecosystem to businesses and take windows out of the equation?

Seriously companies have a ton of systems in place to use iPads/iPhones apps, servers.. I don't think it's unrealistic to think Apple has a real chance at taking chunk of the corporate windows marketshare with this arm based mac.. Though exchange would need to likely stick around.

----------

I don't think Apple would have trouble getting developers onto arm desktop based professional software, any developers knows an apple ecosystem is full or profit potential.

It would not happen to all macs at once, giving time for ecosystem to adjust

 

[Macist]

ARM are a design house that make incredibly advanced processors. They have recently created server class chips.

Judging ARM by their mobile chips is stupid. If there was a big enough customer for laptop or desktop class processors they could create something superior. It's just better technology.

I doubt Apple will go for universal binaries this time. Probably they'll introduce an MacBook Air with incredible battery life together with all their Apps recompiled for ARM. As third parties recompile theirs they'll slowly transition.

 

[winterspan]

It is clear to me that an ARM-based laptop will happen. Timing is everything and will only be introduced when it is ready. In this date and age, it doesn't make sense to have laptops that still get so hot. ARM solves that issue.

While I appreciate all of the progress apple has made with their custom ARM chips, the architecture itself is no panacea. If they wanted to reduce heat and power, they could easily have used the new quad-core intel atom CPUs, but they haven't because the performance sucks compared to Haswell.
It's going to take a while before ARM could be competitive with Intels best, but perhaps they'll market it as a lightweight, netbook-style machine for light duty.

 

[Mac Heretic]

I find the idea of investing in a new Mac and (all of) pro software almost impossible. Last time, it was a painful process, not going to do it again, really. ARM Macs are not for me, because they ignore an idea of continuity and step-by-step process. Just like Windows RT, which has flopped big time.

The situation is even worse today than it was with PowerPC to Intel -transition, because software business sucks nowadays, even for OSX and iOS. It is greedy and thinking only of cutting costs. Software developers only support latest OSes, iOS 7 and Maverics for their new versions, to cut development costs, and their business models are "creative".

 

[junctionscu]

There is no secret sauce in the ARM architecture that would allow it to beat Intel on power/performance in notebooks. If there was, other PC makers would have released ARM-based machines years ago.

When you scale up ARM performance to laptop class, you take a hit on power. I do not see any advantage (other than price - a big one) to moving from Intel to ARM-based chips. Many disadvantages and consumer confusion. Like others have said, it would have to be a new kind of product (compete against Chromebooks?).

Apple does not differentiate on what's inside the machine. Sure, it has to be competitive with the market, but the reasons people buys Apple products are industrial design, software, and the Apple ecosystem (iTunes, apps).

A big reason I switched to a Apple is because I could dual boot windows (and run it in a VM at basically native speed). As much as I like OSX, most of my work still requires windows, so if they switched to ARM, I'd be switching back to PCs

If Windows RT was a huge success, I could maybe see Apple make the switch, but if Apple moved to ARM in Mac products, they lose Windows and the people who need VM or Windows boot.

This should be an obvious move. Look at the benchmarks and their progress. We'll use geekbench for simplicity, not perfect, but it's well rounded. The A7 scores about 2,600, double the A6's 1,200-1,300. Even if Apple did nothing to improve the ARM architecture, the A8 is expected to have a quad core with the die shrink to 20nm. Doubling the die size to fit the power availability in a laptop would yield 8 cores. 8 x 1,300 (per core) is about 10,400. That easily beats the performance of the current Retina Macbook Pro with about 7,000. If they double the core clock to 2.6Ghz, like some rumors say they have, they could beat the Core i7 with the same number and only 4 cores, giving good single-threaded performance. That's all assuming Apple hasn't done anything to improve their architecture, which they most certainly have.

Intel's days are numbered at Apple.

Performance per core does not scale linearly. If you did have a geekbench score of 10,400, the battery life of that device would be terrible. Once you scale performance up, battery life suffers at a quick rate. There is no secret in ARM architecture for battery life. ARM was optimized for low-power devices, Intel for high-performance ones. If you scale ARM up for performance, you take a hit on battery. I don't see the rationale for adopting an ARM architecture from a technical perspective. It has to be all about cost.

 

[JoEw]

ARM are a design house that make incredibly advanced processors. They have recently created server class chips.

Judging ARM by their mobile chips is stupid. If there was a big enough customer for laptop or desktop class processors they could create something superior. It's just better technology.

I doubt Apple will go for universal binaries this time. Probably they'll introduce an MacBook Air with incredible battery life together with all their Apps recompiled for ARM. As third parties recompile theirs they'll slowly transition.

Exactly. No rush on transition, but the fact it could run ipad/iphone keeps actually gives it more software than current macs.. I think this opens the door for all those iOS developers to get into the mac OSx platform.

iOS is basically a touch based version of OSX the processor is the only thing that divides them and method of input. Imagine just a fraction of those 100,000 apps get optimized for mouse and keyboard input. Seamlessly integrated into ipad/iphone variants as well.

 

[sammich]

Didn't see this mentioned anywhere: the source translation says '4-8 quad core processors' which translates to 16-32 cores.

9to5mac says it as 'The iMac and notebook are both said to have “4 or 8″ quad-core arm64 processors, while the Mac mini has only four.' Which would mean 4-8 total CPU cores.

So, which is it? I'm going to lean on the conservative side and say that there aren't 16 A7/8 cores. Each core is capable of 1300+ in geekbench, so...I've done the math. Apple is not putting a Mac Pro rivalling CPU in the MBA.

 

[mateus]

Apple have their roots in RISC processors, so it's interesting that they might moving back to the 'good old days'.

Plus, ARM was created in by a British company, so I'm happy to wave the flag for Britain even though Apple produce these ARM chips themselves. (Probably wise not to involve the British in manufacturing as the chip might leak oil and fall it bits like a Mr Potato Head after a week)

 

[viizi]

Such a hard transition. At least with the transition to intel, intel already had a significant market share. There is no desktop applications that use A7. Apple must have something planned.