Intel's Broadwell Chips for Most Macs Not Shipping Until Early to Mid 2015

[Steve121178]

Fallacy of sunk costs. The real cost of Broadwell is the 14nm process. If that's working there's no reason not to skip Broadwell and get your schedule back on track by going directly to Skylake. Having a short Broadwell lifecycle and a split Broadwell/Skylake lineup is bizarre and likely a dud sales-wise, not to mention expensive.

Intel call the shots, they don't have to and won't skip anything.

PC & tablet manufacturers have been designing & testing Broadwell specific devices & tech for nearly a year. They would all be thrown into turmoil if Intel suddenly axed Broadwell and skipped straight to microarchitecture that no one has even received samples of for testing!

 

[dragje]

This article is not about the processors that are going to be used in Mac Pro's.

Without processors no macs.

 

[allanfries]

Yeah, aside from those with 3-4 year old Macs......
I just don't get why people seem to think they should upgrade their hardware every year? I mean, Is it a competition?

 

[wizard]

I'll look forward to pulling this quote out in a few years. By making an overarching statement like this, you're going to look like one of those that says "technology will never get to the point where ___". I mean, the fact that you say just shows how limited your vision is.

Terribly limited! I see no reason why Apple could produce a Mac OS based system that ran IOS apps in windows. Further all of the complaints about ARM performance are unfounded. Apple is getting extremely good performance out of an ARM chip designed for low power performance. Relax a few constraints and I'm pretty sure Apples ARM architecture would hum along nicely. That doesn't even get into what a major overhaul could do for Apples architecture.

To put it simply there is no good reason to dismiss alternatives to Intel.

 

[wizard]

My point is very soon an ARM processor will be able to run OSX, with the levels of multitasking most MBA entry level users require. its not a backward step, and it could result in increased battery life or lower cost.

Running a browser, lync and a mail program 90% of my day is hardly CPU intensive.

Architecture wise Apples A7 is almost there already. Trivial improvements such as faster RAM and SSD access would go a long ways to providing acceptable performance. Add more cores and a multi tasking OS will run fine.

As it is though iOS is a variant of Mac OS so A7 is already capable of supporting Mac OS. It is really only a matter of performance something that can be dealt with by minor improvements to architecture and a doubling of clock rate. Apple was very honest in describing A7 as a desktop archtecture, it is a very capable design!

 

[Steve121178]

Apple was very honest in describing A7 as a desktop archtecture, it is a very capable design!

...yet a trillion miles away from Intel's Core CPU's.

 

[phoenixsan]

Ok, so.....

the rumor mill have now some fixed dates..........Seems to me then, that Apple can do a refresh in the mac Mini line in this year. I had the wrong perception that processors fitted for the mini are not in place yet........So I adding to the speculation about the no Mini refresh. Why is Apple not doing it?.....

Other view from this news is about the delays in chip production. It is obvious that this issue affects directly Apple update path. Not so clear if Apple would be holding enough on Intel about their chips plans....After all, it is famous the displeasure the late Mr. Jobs got when the PowerPC consortium never delivers the 3.0 Ghz chips for the old PowerMacintosh.....

 

[theluggage]

I agree it's time to go ARM. In the end, and considering both the latest performance gains and battery efficiency, it seems RISC was the good design. Intel had success with CISC just because of their marketing strategies, but it seems RISC is going to win the battle in the long term.

Intel had success with CISC because of the history with the IBM PC and DOS/Windows left them with a huge x86 legacy software base. The original ARM 2 & 3 chips were desktop processors that could open up a can of whoopass on the 286/386s of the day - but they couldn't run Windows (well, they could - I remember running Windows under software emulation on an Acorn ARM 3 system - but it wasn't exactly fast). The reason that there are currently no ARMs that compete with Intel's desktop chips is, mainly, because they have been unassailable, and ARM have concentrated on the low power mobile/embedded market.

Intel couldn't even compete with themselves when they tried to produce a non-x86 architecture (Itanium).

Also, modern x86 architectures have taken a lot of cues from RISC architecture. In simplistic terms, a modern x86 can be regarded as a RISC core sitting behind a x86-to-RISC translator.

If Apple were to jump to ARM, the big advantage would be neither raw power, nor power consumption/heat, but that ARM's licensing model means that they can pick'n'mix 'building blocks' to produce a system-on-a-chip that exactly meets their needs, and then produce it 'themselves'.

Certainly, to start with, ARM wouldn't be viable in the 'Pro' Apple's pro computers because of the lack of raw speed and the need for x86 compatibility. In the past, with 68k-to-PPC and PPC-to-x86, the switch has been accompanied by a big jump in raw speed, making software emulation viable as an interrim solution. x86-to-ARM wouldn't have that advantage. Things like MS Office and Adobe CS would be the sticking points.

However, I'd be surprised if Apple hasn't already got iWork/iLife, Safari etc. running on OS X ARM, and much of the newer software in the OS X App Store is probably just a re-compile away from running on ARM. I suspect that ARM - coupled with suitable GPUs, i/o controllers etc. - would be perfectly viable for an entry-level MacBook Air.

 

[thunng8]

...yet a trillion miles away from Intel's Core CPU's.

That's a bit exaggeration ..

If talking about core performance, a single core of the A7 at 1.4Ghz gets ~1500 in Geekbench 3... and the A8 is only 2 months away and likely to get much higher numbers as it is a major process shrink as well (going from 28nm to 20nm).

A single core of Haswell i7 2.8Ghz in the 13" MBPr gets ~3300 (64 bit result). Only a bit more than double the performance and uses much more than double the power.

----------

What do you even mean "if"? I can tell you right now it won't, it won't even be close.

I bet that the Apple A8 will compete very well with the Broadwell Core M on performance and will use less power.

----------

The OP said AMD, not ARM. The effort to port wouldn't be anything like that from PowerPC to Intel.

There is no porting involved except drivers for the operating system to recognize the CPU. It is even trivial right now to hackinstosh a PC with an AMD processor.

The current instruction set Intel uses for 64 bit was designed by AMD (AMD64).

 

[Mr. Retrofire]

...Christmas...
Hopefully the logic board on my 2011 MBP survives until then.

If you use the iGPU via gfxCardStatus, then yes.

 

[MarkusL]

Fallacy of sunk costs. The real cost of Broadwell is the 14nm process. If that's working there's no reason not to skip Broadwell and get your schedule back on track by going directly to Skylake. Having a short Broadwell lifecycle and a split Broadwell/Skylake lineup is bizarre and likely a dud sales-wise, not to mention expensive.

While I agree with you that the sunk cost in itself is not a good reason for not skipping Broadwell, there could be other real reasons not to skip it. As you say the cost of Broadwell is the 14nm process, since the die shrink is really the only new thing in Broadwell. But the 14nm process must happen with or without Broadwell, because it is needed for Skylake and the way forward. So once 14nm is here, then Broadwell is done and Intel can start cashing in on it.

For Skylake, 14nm is just one of the pieces of the puzzle. If Intel slashes Broadwell and there are other delays (e.g. architecture development) in Skylake, then Intel risks sitting there with a working 14nm process and no products to sell from it.

 

[nocturnum]

Yeah, the generation of replacers. Damn Apples not fast enough for instagram.

Screw Intel. They surely don't get it, how badly all these fanboys need something new in their otherwise empty lives.

 

[wizard]

of course, but they also are driven to compete by market conditions.
in a market where they have little competition, I believe as others have said, that INTEL are probably holding back or at least not pushing the boundaries..... why should they???

You do realize that this is baloney. Intel has taken a significant hit in the last two years leading to layoffs and underutilized plants. The competition wasn't AMD unfortunately but rather the willingness of consumers to spend their electronics budgets on gadgets that don't use Intel parts. That is why we see Intel trying to creat and go after the 2 in 1 market, the tablet markets. The delay for desktop chips is directly related to getting viable low power products out the door.

My desire is to bring back the incentive..... I think with a competitor, and I doubt AMD's APU is really going to cut it, they would be doing more.

I see a lot of BS on these forums about AMDs APUs but they are significantly better than Intel's chips for casual users.

With the size of the A series chips now and the internal teams apple is building with chip expertise, I dont see why an A series chip could not be used to drive a MBA to the same levels of performance seen today, but with better battery life of reduced cost to Apple (and hopefully end users).

I have to agree here. It wouldn't take much for Apple to enhance A7 to make it more viable for laptop use. Frankly it is know different than Intel altering chips for different markets. Why people don't see this is beyond me.

I have little desire to put IOS apps on my laptop, I just think that Apple is skilled enough to pull off migration such that entry level could be A series and Pro-User could remain on INTEL, but an INTEL with the incentive to accelerate its roadmap.

Some things are beyond Intel's control. Beyond that Apple could solve a lot of problem by not marketing an ARM based laptop as a Mac. The last thing Apple needs to do is to confuse consumers.

 

[Mums]

Sadly I was thinking the same. I'm going to hold on onto my refurbished 2011/12 17", which replaced my 2009 17" MBP. Not interested in current offerings.

Yeah - I got an anti-glare 17" MacBook Pro in April 2011 and it is as perfect as the day I got it - never had a single issue with it!

But just bought a $7000 Mac Pro because I need the power, and the 2011 17" MBP can't run/edit 4K video - or the Logic plugins & sample libraries I'm using. Am keeping the MBP though - love it.

 

[wizard]

If they do, Apple could very quickly change over to them. It isn't like it would take much more than some drivers.

Exactly! Frankly AMD doesn't even need to hit 14NM, however with the Global + Samsung alliance I could see 14nm AMD chips by mid to late 2015. Effecitvely they will have caught up to Intel.

----------

From my understanding Intel is an entire generation ahead in it's fab plant abilities with AMD still ramping up 20nm across it's manufacturing abilities.

I'm looking at the new Samsung / Global Foundries alliance that will supposedly have 14 nm shipping next year. Beyond that AMDs processors are very good power wise considering the process they are built upon.

----------

The best of the best from AMD is not nearly as good as the best of Intel, but AMD tends to whack the tar out of comparably priced Intel chips.

I've seen a lot of mis information about AMD in this thread. Often it seems as if people only are willing to look at the extremes. In any event if you are at all concerned about GOU performance AMD APUs are still a far better value than Intel's. With GPU intensive work they often run cooler too.

Now that doesn't mean that AMD's CPUs are competitive but for many users they are more than good enough.

 

[wizard]

People keep saying that, and it continues to be a bad idea?

That continues to be opinion and speculations!

Yes, it is as fast as CPU from YEARS ago. Apple tends to ship computers with current technology, though, so it'd be a major downgrade. Not to mention the fact that the GPU would likely be an even bigger downgrade. There would be no upside.

Im not sure if people like you lack imagination or what but if Apple went this route I would expect an enhanced chip similar to the way the X series processors where an enhancement over the base model.

It is not a Major operation to enhance the memory interface and SSD interface to significantly impact performance. Of course a desktop or laptop machine can tolerate more heat so Apple would be free to crank up clock rate. Simply put you can not dismiss 64 bit ARM based on performance.

It wouldn't even be a given that battery power would be better, considering the fact that you'd be throwing multiple CPU at something that used fewer before. And that's not even going into the logistics of every piece of software suddenly needing to support multiple cores well.

Most software these days do support multiple cores in one form or another.

As for battery power All it needs to do is beat Intels offerings which even with 14 Nm should be a snap. Years ago manufactures where reporting that ARM cores could run in the 500 milliwatts range @ 2 GHz. Cores aren't a significant power problem these days. Cache and real world interfacing is though.

 

[deputy_doofy]

That would make sense if Intel continually pumped out garbage. However, that is not the case. Netburst was only going to be decent at those high clocks, and Netburst was the microarchitecture that Intel had bet with. I am unsure of what you expected Intel to do at that time .. scrapping the product was out of the question, the "fix" was a higher clock rate to make things work as they had hoped.

The P4C 2.4ghz that I had at that time was a very good chip and AMD was doing nothing that gave it a run for the money (unless you consider renaming their chips... that's all). The irony now is that while Intel is far ahead, AMD has picked up the Mhz race with 4+ Ghz chips to try and compete.

Regardless, there is still much room for code to be reviewed and to be made more lean. The point still stands.

I was, and still am, agreeing with you. Pointing out that Intel was leading the way in speeding up inefficiency rather than working on it (back in the P4 days). NOW? The Core i7s are amazing chips. So are the i5s. Not a fan of anything below the i5.

So yes, developers have to step up to the challenge of writing GOOD code. That concept is why developers today can write fantastic games for the Atari 2600. They care enough to actually push the system.

 

[Michael Goff]

Exactly! Frankly AMD doesn't even need to hit 14NM, however with the Global + Samsung alliance I could see 14nm AMD chips by mid to late 2015. Effecitvely they will have caught up to Intel.

----------



I'm looking at the new Samsung / Global Foundries alliance that will supposedly have 14 nm shipping next year. Beyond that AMDs processors are very good power wise considering the process they are built upon.

----------



I've seen a lot of mis information about AMD in this thread. Often it seems as if people only are willing to look at the extremes. In any event if you are at all concerned about GOU performance AMD APUs are still a far better value than Intel's. With GPU intensive work they often run cooler too.

Now that doesn't mean that AMD's CPUs are competitive but for many users they are more than good enough.

That's good news, we could use some competition for Intel.

That continues to be opinion and speculations!

Im not sure if people like you lack imagination or what but if Apple went this route I would expect an enhanced chip similar to the way the X series processors where an enhancement over the base model.

It is not a Major operation to enhance the memory interface and SSD interface to significantly impact performance. Of course a desktop or laptop machine can tolerate more heat so Apple would be free to crank up clock rate. Simply put you can not dismiss 64 bit ARM based on performance.

Most software these days do support multiple cores in one form or another.

As for battery power All it needs to do is beat Intels offerings which even with 14 Nm should be a snap. Years ago manufactures where reporting that ARM cores could run in the 500 milliwatts range @ 2 GHz. Cores aren't a significant power problem these days. Cache and real world interfacing is though.

You say that I'm lacking in imagination. I might be, but all you're doing is speculating. Yes, they could enhance them to work faster. They still wouldn't be able to keep up with the current i5 that comes standard in a MBA.

 

[yellowtruck]

imac 2015

During the PowerPC days they bought in the chips, they didn't actually design the things. Their input on the design was along the lines of "We want it to do this, this and this" and the chipmakers (mainly Motorola and IBM) did what they asked. The reason why this worked was because chips hadn't gotten that complex yet, but when chips got more and more complex Motorola had to drop out and eventually IBM lost interest in providing Apple with chips that could compete with Intel's offerings so Apple had no choice but to go knocking on Intel's door. IBM was able to keep their partnership going for as long as it did by scaling down a mainframe design, however at the cost that it couldn't be put into laptops (Motorola tried making a mobile G5 for Apple, but had to admit they couldn't pull it off).

The reason why Apple is able to make their own chips in the mobile sector and why ARM based mobile solutions in general are so competitive in that market is because Intel pretty much ignored it for years because it wasn't as lucrative as the market for bigger and more powerful chips. Workload wise a modern laptop, desktop or server processor is a quantum leap ahead of mobile chips.

The last thing Apple wants to do is reveal it's "secret sauce" to a third party like Qualcomm, Nvidia, MediaTEK or, god forbid, Samsung. IBM and AMD are the only players out there with experience in non-mobile chips, but IBM isn't interested in consumer chips anymore (rumor says they're considering dropping out of the hardware game completely) and AMD is an x86 company. Not only that, nether company actually has the resources to compete with the behemoth called Intel. AMD has pretty much dropped out of the performance market and is now committed to lower design workload low price and cost solutions called APU's.

So in short it all really comes down to staff and money, which is why Intel and AMD are the only makers of chips that are "above" mobile, but "below" mainframe for almost a decade and why Intel has such an advantage over AMD.

Thanks for taking the time to write that. Just learning about all of this chip stuff and what an "APU" is (what is that again?). I'm one who maxes out power with complex sound design rendering and video editing but can’t afford the top of the line round black cylinder mac thingy. It’s time to buy a new imac. Fortune has not been so kind to me financially and my 2007 imac 27 is broken and I can’t afford a new one just yet. I’m buying one when I can get my apple credit card paid off to zero which will be Sept – Oct. I have no interest in buying an imac costing over 2,600 after taxes with upgraded memory and hard drive when the CPU is already outdated. Now, I’ll have to wait until Feb according to this article and even then Broadwell will be outdated. Not nice at all. Whats worse is, my second Garageband album is locked up on the crashed 2007 imac. Who knew I could compose music and generate a half million views on youtube with my work.I dont even read or write music, all made possible with Garageband! I miss composing and I have grown unhappy with all this chip delay news as my projected purchase time has been pushed back. Composing music is like breathing fresh air. I cant wait to breathe again!
Me signing to music I composed in GArageband
http://www.youtube.com/watch?v=x_Xj...G-r3-OPJtIGdKSh3WI4khZ&feature=share&index=13

 

[holydude]

No, this is Apple we are talking about. It was primarily because of cost & the fact that Intel ship the best processors on the planet for a range of mainstream devices such as desktop computers, workstation class machines & portable computers

The IBM/Toshiba/Sony alliance was a waste of time and billions of dollars. The Cell CPU was a load of crap & developers hated it.

----------



You don't cancel something that has already cost you tens of billions of dollars in development.

Besides, I firmly believe you buy a new machine as and when you need it. The only time it's worth waiting is when a new generation of processors is due imminently and you don't mind waiting upto 4-8 weeks. You also have to remember that Apple typically waits until 3 months after a CPU is released before they make new Macs available. The only exception this time around could be in Q4 of this year when they announce a new MBA with a Y series Broadwell CPU (now branded Intel Core M).

I see a lot of strong words in your post. And you support your claim how ?

 

[ricci]

I was really looking forward to Broadwell Mac Pro! Not to mention MBP as well! Now what I'm I going to do with my Christmas money?

 

[randian]

For Skylake, 14nm is just one of the pieces of the puzzle. If Intel slashes Broadwell and there are other delays (e.g. architecture development) in Skylake, then Intel risks sitting there with a working 14nm process and no products to sell from it.

Unless Intel stopped Skylake development it should be finished already, since the original Skylake schedule had it being introduced this year if I recall correctly. Introducing Skylake next year would give them an extra year of development delay.

 

[MisakixMikasa]

Terribly limited! I see no reason why Apple could produce a Mac OS based system that ran IOS apps in windows. Further all of the complaints about ARM performance are unfounded. Apple is getting extremely good performance out of an ARM chip designed for low power performance. Relax a few constraints and I'm pretty sure Apples ARM architecture would hum along nicely. That doesn't even get into what a major overhaul could do for Apples architecture.

To put it simply there is no good reason to dismiss alternatives to Intel.

Architecture wise Apples A7 is almost there already. Trivial improvements such as faster RAM and SSD access would go a long ways to providing acceptable performance. Add more cores and a multi tasking OS will run fine.

As it is though iOS is a variant of Mac OS so A7 is already capable of supporting Mac OS. It is really only a matter of performance something that can be dealt with by minor improvements to architecture and a doubling of clock rate. Apple was very honest in describing A7 as a desktop archtecture, it is a very capable design!

The A7 benchmarks are comparable with 2009 Core 2 Duo which is 5 years old. ARM is still far away from any Intel chip, even with entry level Core i3. There is no point to make a laptop that has lower performance than current machine and loss x86 compatibility. This makes zero sense.

Unless Apple can design a chip that is leap forward compare with current Core i5 or even Core i7, there is zero sense for Apple to make lower performance machine.

 

[randian]

The A7 benchmarks are comparable with 2009 Core 2 Duo which is 5 years old.

Not bad for a part designed for mobile phones. The point is that Apple wouldn't put an A7/8 their OSX machines, they'd design a new part for much higher power applications. That ARM product would scream. How competitive it would be with a modern Intel CPU I don't know, but it would be interesting to see.

 

[MisakixMikasa]

Not bad for a part designed for mobile phones. The point is that Apple wouldn't put an A7/8 their OSX machines, they'd design a new part for much higher power applications. That ARM product would scream. How competitive it would be with a modern Intel CPU I don't know, but it would be interesting to see.

It is certainly possible for ARM chip reach to x86 if Apple can keep doubling the performance year after year, which is pretty hard to do. The major problem is backward compatibility. Both two platform transition, from 68K to PowerPc and from Intel, there is leap forward in terms of performance. This makes emulation possible without loss performance. However, from Intel to ARM is downgrade in terms of performance. There is no possible way to emulate x86 on an ARM without experience slow down.

If 68K to PowerPC and PowerPC to Intel is necessary, the whole ARM transition is not even necessary and the con outweigh pro.