Except, for the Sapphire display, it was never said by ANYONE but rumor sites like this, that it would have a Sapphire display. So to say it was being dropped, it never was confirmed it was even going to have it.
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 Diversifies Chip Orders for iOS Device and Apple Watch Amid Predictions of ARM Macs
- Thread starter MacRumors
- Start date
- Sort by reaction score
I think they did, but I'm still waiting for those benefits to kick it.
Actually, I'm waiting for them to kick in for me. I think NASA is still using PPC 601 related processes for Mars Rovers (and I'm sure other things); they work and they've got hardened versions of them and they know at this point that they'll keep on ticking (generally). Ford was using them in their cars for a long time too.
Gary
Flawed on multiple levels.
1. At least one of the GPUs shared RAM space with the CPU. (it is an integrated GPU). Two processor packages are probably not going to share RAM. The two GPUs in the current Mac Pros do not.
RAM + ARM cpu and RAM + x86 cpu means higher RAM bill of material (BOM) costs in the system. Apple wants their 30% margin. Driving up BOM costs with little to no value add isn't likely to fly with them.
2. Intel's Turbo (dynamic over/under clocking) and the "race to sleep" modifications of OS X basically get the lower power consumption for low horsepower workloads and high consumption on higher workloads all with just the x86 solution. If throw very low workload at a modern Intel CPU the clock rate (and power consumption) drops.
3. The over major concept missed is that these are System on a Chip ( SoC) solutions. The notion that the CPU only doing the general computations is largely gone in most Mac and all of the iOS systems now. The SoC is CPU + GPU + IOhub chipset ( maybe + RAM + etc. etc. etc. ). Only need one IO hub chipset per PC system. The external ports ( USB , Display , etc ) are hooked to just one. Once start to include SoCs and have multiple USB , Display , Flash , etc. controllers then driving up BOM costs for no good reason.
There are ARM ( or ARM like ) controllers in a mac. They are basically I/O controllers though with fixed tasks. An Ethernet controller with TCP/IP offload typically has a very small embedded CPU inside. Wifi/Bluetooth ... ditto to handle/offload the more advanced functions. SSD drives have a Flash controller which typically an embedded CPU inside.
The application running processor though is typically all by itself or homogenous.
GPGPU programming is heading to a point where CPU and GPU share an address space and to some extent RAM. But that is more a split on extremely parallel processing and more general processing. There is power saving in sending the computation to the processor that can do it most efficiently but flip/flopping on a mix of the same apps not so much.
If there is some super duper upside to having a low power and high power coupled together an Intel solution that merged some Atom cores and a "big" Core i5 core(s) would probably be more effective to OS X than some cross vendor hodge podge. The concept already exists in the ARM world big.LITTLE
http://www.arm.com/products/processors/technologies/biglittleprocessing.php
Right now Intel is concentrating on getting "LITTLE" to work well all by itself. Once more ironed out, it wouldn't be a big leap from them to same thing if Apple (and a few other system vendors ) wanted to place orders for 10+ million x86 big.LiTTLE like implementations. Intel is opening up to doing more custom work combining their processors if customers are willing to pay. On the more big-iron side of their business.
http://www.theregister.co.uk/2015/01/13/amazons_new_ec2_compute_instances_run_on_secret_intel_chips/
4. The final huge flaw is related the immediately above. There is an implicit notion that Intel can't do a low power x86 implementation. That ARM is the only path to lower power. That is deeply flawed.
It wasn't been a high priority for Intel from 2000-2010 but it is now. Intel used to put the "low power" on process technology 1-2 generations back from the latest stuff. Now, Intel is targeting latest process technology for low power implementations. The ARM threat is bigger than the AMD threat and they have adjusted accordingly. ARM isn't magically more efficient in the low power context. There is some overhead in doing the x86 to micro ops translation but as the overall transistor budget goes up and the translations are dynamically cached it is not a show stopper issue.
Head to head implementation competition with Intel .... Apple isn't necessarily going to win that in the Mac space. Intel has a larger group of as good or better designers than Apple does.
Apple can't manage two different motherboards in the Mac mini, and you expect them to manage two different CPUs? Granted, the lack of a quad in the mini is to push the cash bags they call customers into buying MBPs, iMacs and Mac Pros. But having two different CPUs is ridiculous
Against ARM Macs?
Should I assume you never used PowerMacs or PowerBooks?
Seriously, I only saw two features that came out of Apple's switch to x86:
(1) Can install Windows natively, or emulate Windows at a decent speed
(2) Enables Hackintoshes
Everything else is just a compile button.
And if Windows gets serious on ARM, then it really won't matter.
I'm sorry but once they stop making Intel mac, is the day I stop buying them
Should I assume you never used PowerMacs or PowerBooks?
Seriously, I only saw two features that came out of Apple's switch to x86:
(1) Can install Windows natively, or emulate Windows at a decent speed
(2) Enables Hackintoshes
Everything else is just a compile button.
And if Windows gets serious on ARM, then it really won't matter.
besides a a8 score a 3k on geekbench and a macbook air intel core i5 5k is not so far behind.
http://browser.primatelabs.com/geekbench3/search?utf8=✓&q=core+i5
If Apple just needs a cheaper, limited function, laptop (e.g, chromebook competitor ) they can just throw the guts of a iPad into a iPad variant with a keyboard attached. Something like the 360 display hinge, touch screen Windows convertibles that are out there now sprinkled with some Apple design pixie dust.
That seems to be the major refrain about why Apple needs to go to ARM. Cheaper Macs. iOS devices generally already have the "cheaper" angle covered. Some minor expansions to iOS ( multiple apps on screen and more mature apps ) and will cover a pretty large set of use cases of simply cheaper general usage workloads.
I couldn't agree more.
People don't understand that it's more than the CPU.
The current ARM processor, even if it's 64Bit for the iProducts is only dual core.
It doesn't include the necessary peripheral interfaces that the i3, i5, i7 and Xeon processors have.
Ramping an Arm processor to 3GHz a like the Intel processors is more than turning up the clock speed. While Apple can make the ARM faster, they still must conform to the ARM instruction set. They customize the hardware, but they don't control the ARM architecture.
Competing with Intel to make desktop class processors and been horrible for multiple companies, Sun Microsystems SPARC, Fujitsu SPARC, Silicon Graphics MIPS, Sony with the PS, AIM with PPC, etc..... and last but not least AMD.
Intel controls process technology, has prototype fabs, etc.
You cannot compete if you are using TSMC and Samsung because you are at the mercy of their fab capability. Talk about delays.
You want to see Apple throw money down a rat hole? If they try to make ARM processors compete with i7 and Xeon class processors, they will throw money away and to what end? They won't get better performance and independence from Intel doesn't assure delivery of a faster processor any faster than Intel can deliver.
Right now Apple delivers a processor upgrade per year or so.
To move to ARM they would need to deliver low power processors for laptops in addition to iOS devices, desktop class processors for Mac Mini and iMac then server class that can support 6-8 cores per chip in a 12-16 core setup for MacPro machines.
It's just doesn't make sense.
Well they said that he was frequently accurate but not perfect. I don't see a problem with that statement. Were you expecting absolute perfection in his analysis and predictions?
I'm sorry but once they stop making Intel mac, is the day I stop buying them
I think its a really bad idea to drop x86 platform, I can only see bad things from this shift including a more locking down of OS X. Think walled garden for OS X as well.
You don't need ARM processors to lock down the Mac. That's already happening.
----------
None of what you say is true nor makes sense. AMD has already announced their intention to bring x86 and ARM together. There is absolutely no reason why you can not have both ARM and x86 processors together in the same computer. Apple has already implemented heterogeneous software support TWICE.
We don't know that Apple is even going to use the same ARM chips in Macs as they do iOS devices. They might be designing a whole new desktop chip.
I'll remind everyone that Apple wrote a brand new programming language and I'm sure ARM chips are going to tie in with it somehow.
either (3) or (2b) .... lower prices (and/or higher Apple margins ).
The move to x86 meant that Apple could get out of the chipset business. ( Well they could have on PowerPC too but they didn't really persue any other option there effectively).
(4) Multivendor options. AMD may not have won any design wins but they keep trying. Again helps on costs. The fact competing against Intel helps Apple keep Intel in line.
The desktop/laptop/workstation PowerPC volume market never materialized.
Similarly x86 low level driver writers writing for Windows on same arch as low level OS X. Not going to guarantee maximum perhiperhals for Mac/OS X but isn't going to hurt alot either if one developer machine can do both (BootCamp). [ Some overlap with (1) here but users space apps and barrier to entry for dev development is a bit different. ]
(5) shared R&D costs. Generally all classic PC vendors pay into Intel (and AMD) for x86 solution. Part of that money goes into better design , fabs , etc. with the same focus.
Similar is true of ARM and Imagination Tech too. Apple basically follows along with general direction that most of ARM and Imag Tech's customers are going. Apple's licensing fees are only a fraction of the R&D resources being put into the next gen development of the tech.
Theoretically PowerPC was a three way split on R&D with IBM , Motorola, and Apple. That really didn't work out so well because they weren't all pulling in the same direction. Motorola was skewed toward their "home" space of embedded. IBM had big-iron leanings. Apple really didn't particularly want any other PC like solutions to compete with (so off on their own on custom support/IO chipsets ).
Apple taking major component parts from the same competitive market that 90+% of the PC industry was feeding money into helps Apple long term.
[ If Windows started to collapse. Not in the sense of stop growing rapidly but very significant drops in users, then bailing on x86 would make sense. Microsoft may have stumbled for a couple of years but there are signs they are getting back on track in the overall market. ]
And if Windows gets serious on ARM, then it really won't matter.
Windows RT AMD was more a shot across the bow of Intel than it was signaling that Windows was going to dump x86. Intel is pragmatically subsiding Windows inexpensive tablet x86 systems so it largely worked.
[ there is a tension between Microsoft and Intel in lower cost systems as to which one of those two holds onto the better margins as the price point gets more sensitive. ]
Windows NT Alpha/MIPS essentially were similar stabs that didn't have long term 'legs'.
Windows for Phones is probably going to stay serious about ARM. It probably will do some more merging with Window 8,10 etc. but that's separate.
Windows has even bigger x86 inertia than OS X does. As long as Intel and AMD deliver better solutions that will remain the primary target. Intel was a bit slow to respond to the incursion of ARM solution into their primary space but they are fully energized at this point.
It's nothing to do with price, it's to do with power consumption. Compared to ARM processors, Intel processors are power hogs due to their legacy architecture.
Apple must move to ARM or they will be left behind by laptops with lower power consumption. It might not be an issue now, but it will be going forward.
Me working 8 hours a day for 240 days a year for 5 years will net me over $500,000, so I'm not too worried.
Excellent points.
I was merely trying to point out that Apple has had great success with its own ARM designs without having to wait on off-the-shelf options that work. Apple's A-series chips are always the top of the line each year and are heavily integrated maximizing performance and efficiency.
I wouldn't see why a desktop/laptop chip of theirs would be any different provided they work out the difference in platform/arch.
It would be a slow change - obviously not all at once.
Laptops that work for 20 hours before they need charging and fully charge back up in couple of hours are going big sellers.
99% of the users out there do NOT need tons of horsepower.
The average user wants an instant boot machine
* checks or sends emails
* Social media
* Content consumer - videos/music etc
* Makes a speadsheet or word document
* Play a game
The average person is not a serious gamer that needs 1000fps or is a holywood director
The just need a simple power and not a workhorse machine
99% of the users out there do NOT need tons of horsepower.
The average user wants an instant boot machine
* checks or sends emails
* Social media
* Content consumer - videos/music etc
* Makes a speadsheet or word document
* Play a game
The average person is not a serious gamer that needs 1000fps or is a holywood director
The just need a simple power and not a workhorse machine
Haha. MacRumours are a source site, really they've no idea what they are writing. That's why sites like Cult of Mac and 9to5Mac are better because they actually write and publish their own stories.
this is a freaking joke right? This guy is awful with predictions. Just because he says something doesn't mean it will ever happen! Hell I can say that Apple may sometime in the companies future come out with a Smart Car, Smart House, and eventually a fully functional maid robot that will do everything for me, all controlled by siri and voice commands.
There I said it. It's completely true and it will happen. Oh and if I'm wrong, don't hold me to it, but if I'm right, it was all me, i knew it all along and you can't say i didn't tell you so.
Stop quoting this guy, it makes me want to leave this damn site and never come back over stupid ass stories like this.
There I said it. It's completely true and it will happen. Oh and if I'm wrong, don't hold me to it, but if I'm right, it was all me, i knew it all along and you can't say i didn't tell you so.
Stop quoting this guy, it makes me want to leave this damn site and never come back over stupid ass stories like this.
Considering that systems integration is a core part of Apple's business (even more lately), it is a natural extension of what they've been doing. They can create custom versions of processors that fit closely with their software on the desktop (or laptops), just like they've been doing in Iphones.
Relying on Intel is in fact "crimping their style" so to speak.
The way software is developed is become more and more detached from the platform, and Intel is not the most used platform anymore anyway. Times have changed A LOT.
If you rely on off the shelf parts, you save on R&D and integration costs, but lose out on the profit margin, timing, integration and customization. You also lose out in the fact that it is harder to differentiate yourself from the rest of the pack. What's really different between a HP desktop and an Acer or Lenovo one, not much under the hood.
Raw processing performance is not the main differentiator these days, just one of many elements. That alone makes Intel less important to Apple, or anyone, than before.
Especially when you need suggestions for R (or X) rated devices for your Apple products...
It is true that Intel has a larger group of chip designers than Apple. But they both hire from the same pool (especially in Silicon Valley, but elsewhere as well), Apple has acquired several small companies that hire from that pool, and many in that pool don't like Intel's management style. Intel also has to keep a vast portion of their best engineers focused on the products that currently produce most of Intel's revenues, which is mostly hotter chips.
One key difference is that Apple now has tons of data and metrics to help them optimize chip designs for Apple's products and Apple's typical customers, which may lead them down a slightly different path than Intel's need to create chip architectures for the broader general market. This fine tuning is even more important for extending battery life under typical usage than in raw performance on the benchmarks necessary for CPU marketing. e.g. Apple could optimize a chip design to run faster on typical iOS apps, but slower on Android apps at the same power setting as a trade-off, whereas Intel would have to optimize for the latter as well.
Apple has to wait for Intel to release new chips. I understand that.
But guess what.... EVERYONE has to wait for Intel to release new chips.
This isn't an Apple-only problem.
I can see Apple wanting to get a leg-up on the competition by releasing new laptops whenever they want to. But is going ARM or "non-Intel" really the best way to do this?
I guess they have no choice if they really feel like Intel is slowing them down... but it doesn't seem to bother anyone else.
We just saw dozens of new 5th-generation Intel Broadwell laptops announced at CES last week. You can even buy some of them right now. I understand that Broadwell was delayed... but now that it's here it's pretty good.
If Apple didn't have to wait around for Intel... how much better will their products be?
After reading the comments in this thread... I get the impression that ARM chips still don't have the desktop processing power that Intel chips do.
So the question for Apple is... faster chips on someone else's schedule?
Or OK chips on your own schedule?
Ah, cheers Olly I thought they had some sort of link to it. Acorn computers.. I had an Electron, may even still have it! My School had BBC Micros hehe.