Apple wants to move Macs away from Intel chips - report - Page 4

deleted

Edited by MacRulez - 1/22/13 at 7:05am

 

 

Maybe, but hardly worth a thumb-in-one's-own-eye vendetta:  http://news.cnet.com/8301-10805_3-57523672-75/remember-ultrabooks-yeah-no-one-else-does-either/?tag=nl.e404&s_cid=e404

We are seeing with the A6 chip that Apple is leveraging better chip design to produce faster hardware, and I think they will with the Macbook Air to start as no one is using Bootcamp on an Air. Possibly the Mac Mini also. 

 

My question is; when will Apple start to FAB their own chips? Sure, it cost a few Billion to build a FAB and several hundred Million to re-tool, but they have the money and at some point the economics and supply chain advantage will outweigh the cost. How many Ax processors goes into iPods, iPhone, iPads, Airport, AppleTV, possibly the Mac Mini and Macbook Air. That's quite a few million chips. 

I also wonder why Apple could not put a version of the Ax processor into all their Macs for the GPU? I don't think anyone can argue the graphics performance of the A6 and perhaps the A7 will do dual work also using OpenCL.

Edited by Richard Getz - 10/3/12 at 3:00pm

 

Mac Sales have gone up because Windows can run on a Mac natively and so the perception of grand canyon between OS platforms in a physical level disappeared.

 

Adding AMD to the mix [note: not replacing Intel] would be a shrewd move and allow Apple to install AMD APU solutions for the Mac Mini and even offer a BTO option for the Mac Pro that includes the Opteron Line, not to mention the new Piledriver/Steamroller set up with heavy emphasis on OpenCL throughout since AMD is 110% committed to it and Nvidia could give a rat's ass about it.

 

This solution is still 4 years out.

 

Take out the Pro. There is nothing Pro about that set up.

 

Contracts get modified all the time. If Apple wants to use both CPU vendors, Intel isn't going to lose that relationship when Apple could very well buy AMD for a whiff and then Intel is truly in a bind.

 

Thanks, I was just trying to remember what this was called. I think Apple can leverage their A6/A7 processor for the GPU and allow for OpenCL. 

 

They'll never do it on their own, they don't have the IP for it. They'd need to buy a TSMC or GlobalFoundaries to get it, and then managing that part of the company is a nightmare. Plus they'd be stuck with what their own R&D comes up with going forwards. They're in a much better place if they keep the flexibility to play TSMC/GlobalFoundaries/Samsung off of each other for pricing and process performance.

 

That's not how it works. AMD can't currently compete because they A) don't have the 22nm process that Intel is using, and B) made design choices with their current architecture that don't result in better performance for many common applications. Intel has had serious monopoly problems with AMD in the past, and there's just no way they'd ever be able to "tell AMD they can't compete."

Wow. If Steve Jobs were here he'd say, "That idea is a pile of ....."

 

We exist in a world of computers that use both Windows and Macs. One can't be all things to all people. (Very few games on Mac, and a lot of pro software is Windows only, eg)

 

As a consultant, I use the fact that the Mac does both Windows and Mac to sell Apple computers. And I've sold a lot of them. Remember the world still uses Windows mostly.

 

Gee, I wish Steve Jobs were here to bash the idiot who thought of nixing Intel. Stupid in today's world.

deleted

Edited by MacRulez - 1/22/13 at 7:04am

 

To clarify, no reason if the size was bumped up to 1U and kept this design. The design as shown is too small, but that's only because we have an ODD for a frame of reference.

Wouldn't that involve an awful lot of risc?

 

 

😩やれやれ。。。

Moving away from Intel in the near future has about as much chance of happening as Kim Kardashian not being a fat-assed attention whore; which is to say that it's certainly possible, but not very likely at all.

if you want to make Aperture run as fast as possible: you would use GPU.

 

same thing will happened with all power-hunger software ("power-hunger software" aka DESKTOP software: CuBase, Logic, Premier, FinalCut, PhotoShop, 3D Max, AutoCad...) - it will be migrated to GPU, away from x86.

 

so x86 would become irrelevant even in Wintel world (soon or later)!! 

Wait. Apple needs to differentiate their computers from PC's running Windows, and changing the CPU is the way? Ludicrous.

 

Apple can squeeze into the volume market for the Mac Mini with an APU and reduce the form factor while having a 7660 AMD OpenCL 1.2/OpenGL 4.x compliant GPGPU with 384 real stream cores allowing the value for the Mac Mini to go up significantly while the cost drops.

 

Let me know when 7000/600 series AMD/Nvidia half-height single slot GPGPUs with 1/2G/3GB of RAM on one, never mind more than one GPGPU that currently the Mac Pro accomodates with a 1500W Power Supply all manages to fit within a < 400W low profile aluminum chassis with room for multiple SSD drives and not just a Thunderbolt jack to an additional RAID ready 12 TB optional stack, all within that form factor.

 

Oh that's right! They don't exist.

 

Nixing Intel doesn't mean Nixing Intel. It means augmenting Intel with another Intel compliant CPU vendor, AMD. Personally, they should build out models for both CPU lines or target specifically the Mac Mini for the APU line from AMD and allow the big box [Mac Pro] to be BTO for either Intel or AMD. All of Apple's code is optimized either way and Apple doesn't make a dime advertising for Intel.

 

You really don't have a clue the differences between Embedded systems software and Workstation level software with a full OS X stack I see.

 

You think? Your comment prompted me to do a little reading about it and it seems like its only real trick is power management. Granted that's important, but it doesn't really strike me as particularly exciting.

 

Oh, and to those saying that CPU speed increases are no longer as necessary as they once were, may I suggest you spend some time working with the latest Adobe apps? Installing those made my MacBook Pro instantly seem three years older than it is!

 

Huh? Did you miss the part where I mentioned that adding Windows to a Mac was more convenient than setting up another machine to run the Windows side of things?

 

The fact that SOME titles are available in OSX versions isn't particularly helpful if the one you're using isn't one of them.

 

Bad example because yes, it is.

 

A "pro" machine would be serviceable by in-house IT personnel, adhere to generally accepted industry standards for replaceable components, and not make the most widely used software in the world look WORSE, not better.

 

In this case two of those three issues are a direct result of the absurd drive to make an already-super-slim computer less than an 1/8" thinner, so it is less "pro" than previous models just because it's slimmer.

deleted

Edited by MacRulez - 1/22/13 at 7:04am

That's ridiculous. Who told you that no one is using Boot Camp on the MBA? Or virtual machines?

As it is, Windows compatibility is a very important feature. Apple's not going to give it up easily.

More importantly, the A6 is probably 1/10 the performance of the slowest i5 in the MBA. Apple is not going to release a newer product with a 90% decrease in performance. It was bad enough when they had to switch from PPC to Intel - where the Intel chips were comparable (maybe a 10-20% decrease in performance AT MOST, and often much less).

The entire premise of them REPLACING Intel chips with ARM is ridiculous.

I could, however, see them selling a MBA-type machine which is essentially an iPad Pro. That is, an iPad with built in keyboard, larger battery, maybe higher clock speed, etc. It would do everything the iPad does (email, simple documents, light graphics, work, etc) with the benefit of an attached keyboard. But it would be in addition to existing products - it would not replace the MBA.

I think it's worth making a distinction. I happen to be one of the people saying that CPU speed increases are not as important as they once were - HOWEVER, I specifically limit that to the average user. For the average user, incremental increases in CPU performance no longer matter much. I remember the days when even a brand new computer felt slow and you eagerly awaited any performance improvements. Today, I'm using a 6 year old MBP and it's perfectly fine for almost everything I do. As an average user, today's computers are fast enough that the computer is no longer the bottleneck.

However, there will always be pro users (and, to a lesser extent, gamers) who need every bit of computing power available. But that's a relatively small percentage.

First, not very many people buy Macs solely to run Windows. Many people prefer Macs and Mac OS X and buy Macs to run Mac OS but appreciate the ability to run Windows on occasion (for the one app that their IT group insists on which only runs on Windows.

Also, some people prefer the quality and design of Macs. For example, no one offers anything like the rMBP. Until recently, nothing was close to the MBA.

I don't think Apple would partner with AMD for quite a while.  While I believe AMD offers a good price/performance value, I don't think it would be good for the perception of Apples computers' premium branding.  It would be interesting to see how an Apple OS could take advantage of their present architecture, unlike Windows 7.  Supposedly the scheduler in Windows 8 will offer up better performance for the Bulldozer/Piledriver line, maybe Steamroller would be even better, but outside of running a Hackintosh or maybe a Linux flavor it would be tough to get an adequate comparison.  I am not even sure if you can run a Hackintosh on a current AMD chip, but I am sure someone will take up that challenge.  

 

In fairness to Intel (can't believe I just typed that!), they do have other customers to consider.  Judging by the PC to Mac marketshare, the PC world is where they make the majority of their profits, so I could see them trying to get manufacturers to develop those Ultrabooks.  However, some of the external designs of the Ultrabooks are blatant MBA rip-offs, but that's not Intel's fault as far as I know.  I wish AMD could better match the performance of the Intel chips, but with their limited resources, it is amazing they are doing as well as they are.

 

Didn't Intel use AMD tech/IP as well once their 64 bit CPUs came out?  Not sure, but I remember reading something about that years ago.  I thought it was AMD's limited R&D budget that prevented them from competing with the high-end Intel chips, not licensing deals.

 

As you can already get keyboards for iPads in various flavors, I don't see that happening.  However, I didn't see a smaller iPad coming either.

 

Oops, I guess I should be checking back more frequently. Thanks for the compliments! I used modo for that. Also, LOL at the arguments springing up from this...thing I literally just tossed out in a few hours without a single thought about how components would physically fit inside.  I was riffing on the idea from Mrgan's list of nonsense Apple product names, that's all. Take a Mac Pro, make it thin.

 

AMD went with x64, which was backwardly-compatible with x86, at the time when Intel was trying to sell everyone on the Itanium processor.

 

IIRC, AMD still pays Intel licensing fees for the x86 instruction set that x64 was built on.

Yeah - ARM in any Mac computer? No. Arm has only just managed to beat a Single Core Atom - a modified Atom at that to run in a smart phone, no less. To my knowledge they were Single Core Atoms as well - can I see comparisons with the Dual Core Atoms used in net books? ARM as a co-processor, however, I like the idea of that. Use the ARM cores to eliviate stress from the graphics processor or something - or use it to run non demanding background services like managing network connections and bluetooth transfers - or even (in the most old-school of moves) using it as a maths co-processor for floating point calculations.

 

AMD? Possibly - processors like the AMD A8-5500 and A10-5700 are actually not bad - in fact they're actually very good! The A10 in particular can go toe-toe with a desktop Ivy Bridge Core i5 and have far better graphics on chip as well - my friend has an AMD A8 based HP Laptop and it can run many games at playable frame rates and modest detail settings. If a Mac Mini or one of the low end iMacs came out with an A8 and an A10 (respectively) I'd be fine with that. Low thermal output, nippy performance, good power per watt. AMD are also very good at making cheap but efficient ultra-mobile processors as well - things like the Vision E-400 are AMD's answer to the Atom and completely obliterate it in terms of performance (even at lower clock speeds) - a beefed up Vision or a scaled down Ax series would be a winner in a MacBook Air.

 

The higher end iMacs, MacBook Pros and MacPro, however, would need to stick with Intel. Intels Xeon and Core i7 CPUs destroy AMD in every sense of the word. AMDs workstation chip, the Opteron, is just no where near as fast as the Xeon and the Ivy Bridge i7 (hell, even some first generation i7 chips from 2009/2010!) leave AMD in the dust. Macs are premium machines and the top end models are excruciatingly fast that more often than not warrant the premium: even the two year old MacPro with a single Hex Core Intel Xeon W3680 has the same benchmark performance as the recent Xeon E5-2640.

 

 

But shifting focus away from CPUs, you have thunderbolt. Its currently on all recently released Macs and tied to the Intel chipset. Unless Apple can pull some fairy dust out of their backsides and wish upon a couple of stars, Thunderbolt on an AMD platform is likely not going to happen soon.

I don't think you have to worry about it, at least for a long time.  Apple knows if they did make the change, it would have to be seamless.  What would happen if they make the system faster than the fastest PC and still run Windows?  I mean, they were able to run Windows on a Mac before they switched to Intel chips. 

Since none of us readers making comments don't know what is in Apple's research labs, all we can do is speculate on possibilities and see who the closest.

Speculation is fun though!

Here's what I think, realistically, Apple would or could do:

1. Buy S3/VIA/Cyrix, jettison the trash and low-build quality they're known for. May as well not, other than for any x86-64 patent/licences.
2. Buy nVidia, er probably not, nVidia seems to want to do it's own thing.
3. Buy AMD/ATI. AMD doesn't make horrible parts, and they do hold 22-30% of the market, roughly the same amount Apple has for Desktops (not all systems.) The problem here is that this would actually remove the only competition for x86-64 parts, forcing the likes of Dell and HP to only have Intel to buy from.

However there is nothing saying that Apple couldn't licence what AMD/ATI has and improve on it. It doesn't seem far fetched, but moving to ARM seems silly as the ARM parts don't yet compete on performance, only power. The ARM parts are only therefor good in low-power applications (Mobile phones, tablets, and embedded systems.) They are moving into servers, but they're not replacing Xeon's, rather they're replacing on-demand cloud systems (that just run web servers.) eg instead of buying 2 top of the line Xeon E5's for 20,000$, you instead buy 80 ARM cores and instead spread the load among multiple chips. Linux, FreeBSD, and all the open source software like Apache HTTPD and MySQL can run on it. So unless the next Mac Pro has 80 ARM cores in it, I don't see this going anywhere. Likewise the iMac's would probably need like 20-40 of them.

You also can't emulate x86-64 on anything. So it's unlikely that Apple would again change the CPU on a whim, as this would kill sales and spell the end of the Mac. You will more likely see both ARM and x86-64 CPU's in a future Mac, and the OS will switch to the ARM parts when high performance isn't needed, eventually developers will make their code run on both. But something tells me that customers aren't willing to switch from x86-64, as developers will just go back to Windows.

AMD's problem has been that Intel is too far ahead. AMD produces chips that consume more power than Intel's, so you don't get the same performance out of the top end desktop AMD chip that you do from the top end desktop Intel one. However Intel is also failing, as processors released this year Intel Core i7-3770 @ 3.40GHz (10359) is the same performance as processors two years ago: Intel Core i7 995X @ 3.60GHz (10374). AMD's fastest desktop part is still 20% slower: AMD FX-8150 Eight-Core (8172) which is comparable to Intel Core i5-3570K @ 3.40GHz (7726)@ http://www.cpubenchmark.net/cpu_list.php

But also, AMD's FX-8150 is 125W @ 32nm while Intel's 3rd generation i5 and i7 are 77watts @ 22nm vs intel's first generation i7 995X at 130watts @ 32nm.

So in a sense AMD is 2 years behind Intel. Apple couldn't use AMD's parts since it would offer less than what is currently being offered. Large data centers and Enterprise will pick parts that are most power efficient, and AMD is losing here.

A move away from Intel chips is not unthinkable.  The PPC architecture had its devotees.  But IBM and Moto's stagnation brought about its end.  Politics, I guess.  Plus Apple's marketshare probably couldn't justify the investment required to take on Intel.

 

Now?  Matters changed somewhat.

 

Mac sales have torpedoed from less than 1 million per quarter to 4 million to right past it...to 5 million sales per quarter give or take a few hundred thousand...and still, just about...growing...though it has slowed up significantly.  We may find out soon what the head room is for a 'premium' and 'cool' computer is.  Of those 5 million?  It's laptops 4:1.  The laptop is the 'new' desktop.  And you can sit it on your desktop...your laptop...in your bedroom...on the go.  I don't like them.  But I recognise why they're popular.  They're so much more powerful now than when I first took a dislike to them.  They're sleeker and more powerful, have far more powerful gpus than they used to, waaay more ram...and...far bigger HDS...and faster ones(!) if you include SSD...not to mention the performance of the i7 bringing quad cpu to the laptop.  Given all that?  All the pre-eminent advantages of the desktop have been swallowed by the laptop.  And sales reflect that.  Instead of the clunk of 'most' desktops...people are voting with their wallets.  The current Macbook Pro retina is a work of art.  It's the best Mac they make.  By a mile.  It's superb.  And the screen makes the 27 inch iMac blurry.  Sorry, it does.  And I'm an iMac owner.  But I have to hold my hands up.

 

So, what advantages does the desktop still have?  Well.  Bigger screen for starters.  But the laptop can 'dock' at a 27 inch screen?  And...faster cpu?  (But not by much...)  ...faster gpu?  (Not by much...)

 

So.  I can see why laptop are super popular.

 

That doesn't stop Wizard and myself wanting our dream Apple desktops though.  The current frankenstein desktop range by Apple is far too extreme...a richness of poverty.  The mac pro is a joke in price and specs.  The iMac is over a year out of date.  15 months?  Shocking.  (Has it really taken this long to develop a new one..?)  and the mini is a over priced sawn off biscuit tin.  (Though sexy biscuit tin...with no k/b, mouse...groans...etc.  no screen...bundled...groans more...and priced over £500 just for that?  Give me a break...all the reasons why some people won't buy Apple and I can understand why....)

 

So (my 3rd 'so'... :D) what does this have to do with moving away from Intel.  Quite a lot I should imagine.  The vast amount of cpus are now Arm style in terms of Apple's computer business (and for those in denial the iPhone 5 - *waves at you know who... and the iPad and the iPods are computers...and more powerful computers than the old G4s from a mere decade ago.  Times have changed...)  We're talking how many cpus?  25 million iPhone 5, 5, 5, 5, 5 (sorry, can't resist saying '5') cpus?  15 million iPad cpus?  How many millions iPod cpus?  What we talking here?  50 million cpu sales per quarter of the Arm variety?

 

Now.  Let's look at the Mac sales (though impressive in context of Mac lore...)  5 million give or take?  Out gunned by a factor of 10:1!!!  

 

Looking at the laptop to desktop sales?  Wow.  But looking at the ratio of Apple's revenue in terms of volume and profit of Arm/iOS to Mac/Os X?  The writing is on the wall.

 

As soon as an Arm chip, 64 bit et al is good to go?  We can expect Apple to begin seriously moving away from Intel.  Not ready yet.  But it's inevitable.  The evidence is there in the iPhone?  The iPod?  The iPad?  3 big clues right there.  We're already IN the transition.  Look at the 'iOS' 'integration into Mac Os X.

 

SoC is where we're headed.  Big mainframes of massive computers that send man to the moon (allegedly...though I still want to know more about how they got past the radiation belt and that coke bottle...and why we've not been back there since...) are a thing of the past...  It's all about the tiny.

 

iPhones can murder G4s.  Progress.

 

The sheer volume of ARM and power (these things are really surging forward in terms of cpu processing and gpu processing.  The lastest iphone 5 looks just about PS3 class in graphics.)

 

Intel?  Look in trouble to me.  If not now, in the future.  The Wintel hedgemony faces it's first real threat as iOS, Android, tablets, phones hoover and demolish new Wintel sales.  M$ stumbles with Windows released.  And more people have 'enough' computing power for their needs these days.  Which again reflects in the iMac to Pro sales...Prosumer vs 'niche' and Laptop to iMac sales....and in iPad to Laptop sales.  Ah.  We see a pattern.  There's a sliding scale of 'good enough' and Jobs was right.  For 9/10 things?  A tablet is good enough and going to get even better!

 

It has a retina screen and with the iPad 4?  Set to get even more powerful in cpu and gpu power.  How long before you can 'dock' it on a 27 inch retina screen?  It's already got a higher resolution than than most iMacs?

 

Will we have ARM chips capable of running Photoshop?  Lightwave?  Video?

 

Well.  We have iPhones and pads that can already image edit, edit hi def video, play games (the g4  could only dream about...) in apps that have a far less bloated and efficient footprint.  Hundreds and thousands of apps in a far competitive market...to dwarf the Mac market and eclipsing even the Wintel market now.  I have one app on my iPad that is super fast.  It would make Corel's Painter blush.  It has layers and brushes and no interface or feature clutter.

 

An Arm future?  It's already happening.

 

What does this mean for desktops?  Well.  Look at the drawn out update schedule.  From the Pro to the Mini.  If I was Apple?  Why not just make one box.  A cube-esque shape that can take 'more' powerful components...than a less than 1 inch laptop.  Simple.  From integrated crappics to mainsteam to workstation.  

 

Done.  They get to steamline the desktop model to just 'one.'  They let customers buy the cpu/gpus they want within a limited choice.  ie am I a light, prosumer or workstation user.  Job done.  Price can reflect that.  Plug it into the 27 inch retina display?

 

Jobs a good 'un.  This could last until the giant iPads wipe out desktops (or become the new desktops and supercede Laptops as the 'new' desktop...all things in time sort of thing...) or until iPads become powerful enough to be docked on a bigger screen...or adjacent to it...like we have now.  Desktop just becomes semantics in time.

 

Will this happen in the next year?  Probably not.  But the signs are there that we're moving OS and iOS devices in that direction.  An iPad can already do most of the stuff a G4 could do.  I'd expect that accelerate past the Macs of only a few years ago within a year or so.

 

An Arm chip in a Mac?  I wouldn't rule it out.  Apple likes that control.  They have it in all their iOS devices.  They'd surely want it on their Macs or they wouldn't have moved to Intel.  ie to get the power savings to push their hardware designs.  And if Arm becomes powerful to facilitate that?  It's going to (and is) happening.  They're just waiting for the power to weight ratio.  And as soon as that hits, good buh bye Intel.

 

Look at the size of the iOS apps market.  What do you need Windows for?  (Cue example of a handful of apps...that will soon be swallowed by said juggernaut...  Wintel has finally met it's nemesis.  iOS.  It's a deadly enemy that is taking the Wintel 'juggernaut' out...piece by piece...in a pirana Tsunami...)

 

Lemon Bon Bon.