Intel Skylake / Kaby Lake

[R0H1T]

Whoops I was looking at q1 #s.

I don't understand your line of thinking at all. Maybe some made up numbers would help. How much do you think it costs apple to make a $1500 macbook pro? How much of that do you think goes to intel? How much do you think they would save by making their own CPU? What is the profit margin increase on the average mac?

I don't see any reasonable set of numbers resulting in a profit increase that makes the extreme amount of hardware and software engineering work make sense.

Just going by the number of Mac sales, around 4.2 million last quarter & increase of 4% YoY. The iPhone sales went down YoY & I think they've already peaked annually, the decline is less stark atm because more markets, like India, are getting the latest iPhone as opposed to a dated model, also there's growth outside of the traditional Apple strongholds.

As for how much does Apple pay, I'd expect at least a $100 worth of Intel goes inside each Mac, a lot more for Iris GPU parts & likewise the desktop & Xeon parts. The net (%) margins on Mac might be lower than iPhone or iPad but per unit they make a lot more as well. If I take your number $1.75 billion saved, or just a billion for arguments sake, that's still a good billion saved. The thing is Mac aren't slowing down in sales, they're also getting more expensive at retail, whether due to component costs or more profits for Apple isn't clear atm.

As for hardware engineering, that's an ongoing process, the Ax have come from single to 2+2 core config. There's no reason to think that Apple won;t go full quad or octa core next time, an Ax SoC already has the highest IPC in it's TDP range, possibly even higher than core m. What makes you think that Apple won't go for more software integration ~ http://www.macworld.com/article/3120857/macs/macos-sierra-review.html

A large part is already underway, it's just a one time effort not unlike the last transition from PPC.

 

[TheF34RChannel]

https://videocardz.com/71269/intel-core-x-series-full-specs-revealed

Full Spec list for Skylake-X. The 18 core has a base clock of 2.6 with TB3 of 4.4 while having a TDP of 165.

I thought I put the screenshot in the thread for some easy and quick viewing:

 

[PeterScott]

Just going by the number of Mac sales, around 4.2 million last quarter & increase of 4% YoY. The iPhone sales went down YoY & I think they've already peaked annually, the decline is less stark atm because more markets, like India, are getting the latest iPhone as opposed to a dated model, also there's growth outside of the traditional Apple strongholds.

As for how much does Apple pay, I'd expect at least a $100 worth of Intel goes inside each Mac, a lot more for Iris GPU parts & likewise the desktop & Xeon parts. The net (%) margins on Mac might be lower than iPhone or iPad but per unit they make a lot more as well. If I take your number $1.75 billion saved,

It's not a savings that can be booked. I would argue that a Mac that can't run parallels/Wine for cross platform software is a less valuable mac. A mac that can't run legacy Mac software is a less valuable Mac.

IMO, you end up giving those "savings" to customers, and you would lose sales as well.

Switching Macs to ARM is Lose-Lose-Lose-Lose.
1): Hundreds of millions wasted in transition.
2): Painful transition at a time when desktop sales are shaky.
3): Less useful, less valuable Mac selling for less.
4); Less useful, less valuable Mac selling less volume.

Fools errand.

 

[R0H1T]

It's not a savings that can be booked. I would argue that a Mac that can't run parallels/Wine for cross platform software is a less valuable mac. A mac that can't run legacy Mac software is a less valuable Mac.

IMO, you end up giving those "savings" to customers, and you would lose sales as well.

Switching Macs to ARM is Lose-Lose-Lose-Lose.
1): Hundreds of millions wasted in transition.
2): Painful transition at a time when desktop sales are shaky.
3): Less useful, less valuable Mac selling for less.
4); Less useful, less valuable Mac selling less volume.

Fools errand.

Not when windows will run on ARM, or did you forget that? The parallels/wineverse won't do better than MS's own x86 to ARM translation, also this ~ https://www.techpowerup.com/234185/...tel-and-qualcomm-microsoft-over-x86-emulation

If Intel ever did the unthinkable of suing MS then Mac isn't the only turf they'll have to defend.

 

[PeterScott]

Not when windows will run on ARM, or did you forget that? The parallels/wineverse won't do better than MS's own x86 to ARM translation, also this ~ https://www.techpowerup.com/234185/...tel-and-qualcomm-microsoft-over-x86-emulation

If Intel ever did the unthinkable of suing MS then Mac isn't the only turf they'll have to defend.

Windows doesn't run on ARM. This is intended to be a special case for Qualcomm chips.

 

[R0H1T]

Windows doesn't run on ARM. This is intended to be a special case for Qualcomm chips.

Just for QC right now, if Apple pays or makes a compelling argument to MS then why wouldn't there be a bootcamp for the new Mac OS?

 

[PeterScott]

Just for QC right now, if Apple pays or makes a compelling argument to MS then why wouldn't there be a bootcamp for the new Mac OS?

It runs no where for now. If/when it finally ships we can finally see how bad an idea emulating x86 on ARM really is.

 

[R0H1T]

It runs no where for now. If/when it finally ships we can finally see how bad an idea emulating x86 on ARM really is.

Yes, btw same token your QC is a special case is just that an assumption. We'll see if or when Apple wants to make more money by turning inwards, till then we'll see what unfolds slowly.

 

[Zucker2k]

The actual term is instruction latency, measured in cycles, which is the analogue of time when it comes to computing. A faster CPU would have lower latency executing a given instruction, hence the more number of instructions that can be run in a given number of cycles. Equalizing the clock frequency is one way of ensuring that you are considering the same number of cycles.

Why don't we equalize clocks between Ryzen and SKL-X at 4.6Ghz on water cooling then? It won't be a fair fight, right? I hope you see that even the decision to run clocks at 3Ghz is an indirect recognition of the limitations of one design compared to the other?

 

[R0H1T]

Why don't we equalize clocks between Ryzen and SKL-X at 4.6Ghz on water cooling then? It won't be a fair fight, right? I hope you see that even the decision to run clocks at 3Ghz is an indirect recognition of the limitations of one design compared to the other?

You forgot another key aspect, process ~ http://www.tomshardware.com/reviews/overclocking-amd-ryzen-ln2,5116.html

I for one believe Ryzen is severely limited by GF, a better foundry might've net them half a GHz or more in terms of OC headroom.

 

[Phynaz]

The day MacOS is ported to their Ax SOC, and Macbooks are equipped with them, is the day Intel needs to be worried about.

I'll bet MacOS has been running on ARM for a long time already.

 

[DrMrLordX]

Until what? Apple starts selling CPUs to third parties? iPads replace everyone's laptops? iPads weren't doing so well the last time I checked.

All it takes is one company demonstrating that Intel can be beat in a consumer-level (or worse, professional-level) power envelope. The rest is too difficult to predict, but the core of Intel's dominance has less to do with market position or even software ecosystems than you might think. It's the notion - no, the fact - that it's impossible to get anything higher-performance than an Intel-based system for general computing needs.

If people see better perf/watt and cost/node metrics from some other company, they'll transition over.

Also consider how Intel got their foot into the enterprise computing door years ago with the Pentium Pro. They knocked off a bunch of established players selling Big Iron by showing people how clusters of commodity hardware could replace the "supercomputer" paradigm. Today Intel produces mega-multicore chips that burn over 100W per socket for server farms. Now imagine something like a modern or next-gen Apple SoC running @ 3 GHz in a cheap scale-out board with good/great memory-per-node. The first thing Intel will lose is the huge VM market where most VMs are going to be allocated to 4c/4t or 4c/8t anyway.

The day MacOS is ported to their Ax SOC, and Macbooks are equipped with them, is the day Intel needs to be worried about.

More importantly, the day that Apple and its *BSD devs start porting a bunch of their software to Ax is the day Intel needs to show fear.

Switching Macs to ARM is Lose-Lose-Lose-Lose.

It didn't stop them from switching to x86 years ago. Ultimately they went with the stronger hardware platform, and it worked for them.

 

[PeterScott]

It didn't stop them from switching to x86 years ago. Ultimately they went with the stronger hardware platform, and it worked for them.

Moving from PPC-x86 was pretty much a slam dunk. Moving from low volume to the market leader makes sense.

This would be moving from market leader to low volume parts, and needing to design a whole new line of chips. It's not the same case at all.

If Apple is looking to save a few bucks on parts cost, they have a MUCH simpler option: AMD.

Really this is way off topic. There must be a Apple-ARM thread somewhere if you want to keep discussing this.

 

[DrMrLordX]

Moving from PPC-x86 was pretty much a slam dunk. Moving from low volume to the market leader makes sense.

This would be moving from market leader to low volume parts, and needing to design a whole new line of chips. It's not the same case at all.

Who said anything about low volume? And Apple is ALREADY designing new chip after new chip, making great leaps and bounds as they go.

If Apple is looking to save a few bucks on parts cost, they have a MUCH simpler option: AMD.

They have a better core in-house than anything Intel can bring into their space. Why waste time with AMD?

Really this is way off topic. There must be a Apple-ARM thread somewhere if you want to keep discussing this.

The topic is: why does Intel have to make an innovative and dramatic jump in uarch performance over Core? The answer: Apple has already beaten them in IPC. Intel needs to hit back with Icelake. It has very much to do with Intel, and the fact that Coffeelake has to be their last iterative improvement in the Core line. If they can't grow beyond that, they're in trouble.

 

[moonbogg]

2.6ghz base on the 18 core. I absolutely cannot wait for people to get their hands on that chip and try to overclock all those cores past 4ghz. Its going to be beyond interesting to see how the platform behaves with so much juice being pulled. Its going to be insane no matter what happens. Insanely fast, insanely power hungry, insanely explosive. All of the above are welcome.

 

[PeterScott]

Who said anything about low volume? And Apple is ALREADY designing new chip after new chip, making great leaps and bounds as they go.

They would need multiple new chips for Macs, those would be low volume. iPhone chips won't replace desktop chips.

They can design chips for iPhone no problem because the market is 10x larger, and uniform (one new chip for each new generation).

Macs with 1/10th the market, and a much more fragmented lineup needing different chips for different segments would make chips even smaller in volume than the 1/10 th size vs iPhone.

 

[raghu78]

2.6ghz base on the 18 core. I absolutely cannot wait for people to get their hands on that chip and try to overclock all those cores past 4ghz. Its going to be beyond interesting to see how the platform behaves with so much juice being pulled. Its going to be insane no matter what happens. Insanely fast, insanely power hungry, insanely explosive. All of the above are welcome.

http://www.anandtech.com/show/11550...-core-i9-7900x-i7-7820x-and-i7-7800x-tested/8

Given that 7900X is drawing close to 150w we can expect a 7980XE at 4 Ghz across all cores to draw 280w or more. Bigger dies are generally more difficult to clock as well as smaller dies especially when pushing all cores to max. So I would expect power draw to exceed 300w. The power draw will be very high in applications which max out CPU usage like rendering or video encoding and lower in games and other apps which don't max CPU usage.

 

[Edrick]

It didn't stop them from switching to x86 years ago. Ultimately they went with the stronger hardware platform, and it worked for them.

Exactly. PowerPC was the better CPU, but Intel had the better overall platform.

When you consider that there are as many, if not more, ARM devices out there than x86 devices....you have to really think about the future.

 

[jpiniero]

http://www.anandtech.com/show/11550...-core-i9-7900x-i7-7820x-and-i7-7800x-tested/8
Given that 7900X is drawing close to 150w we can expect a 7980XE at 4 Ghz across all cores to draw 280w or more.

Using AVX-512, yeah. Non-AVX should be okay for 4 Ghz with all cores.

 

[Edrick]

They would need multiple new chips for Macs, those would be low volume. iPhone chips won't replace desktop chips.

They can design chips for iPhone no problem because the market is 10x larger, and uniform (one new chip for each new generation).

Macs with 1/10th the market, and a much more fragmented lineup needing different chips for different segments would make chips even smaller in volume than the 1/10 th size vs iPhone.

They would not need that many chips for Macbooks. They already have the A series and Ax series. Adding one more to the lineup would not be that difficult (think an 8 core A11X for example). And since they are already designing the chips for the iPhone market, it would save them money to use that same technology and R&D in the Macbook lineup. A hell of a lot cheaper than buying from Intel.

Also, they would have full control over their hardware. Something Apple wants to have. Right now, Intel is the hurdle they need to clear in order to make that happen.

 

[Edrick]

2.6ghz base on the 18 core. I absolutely cannot wait for people to get their hands on that chip and try to overclock all those cores past 4ghz. Its going to be beyond interesting to see how the platform behaves with so much juice being pulled. Its going to be insane no matter what happens. Insanely fast, insanely power hungry, insanely explosive. All of the above are welcome.

Going to be just as interesting watching people trying to OC a 16 core Threadripper to 4Ghz. As I said before, I think frequency expectations of these new "many core CPUs" are far too high....from both camps.

Just give me a Power 9 CPU that can run Windows

 

[WhiteNoise]

personally I do not see the need to OC these higher core chips. If I get my hands on one I'll leave it be I think.
AMD breaking into the server world now might work for some but adoption might take awhile. The company I work for buys from DELL and HP but mainly DELL and we buy/sell thousands of workstations and servers a year used in public safety. People I deal with are just comfortable with Intel. They works great, they offer long life and great performance. prices on the other hand are quite high depending on the chip(s) being used. I cringe sometimes when I see the price WE pay and then my ass puckers when I see the mark up to our customers. (Our customers are all the big telecom companies such as Motorola, Verizon, AT&T, Frontier etc)

My point I guess is I know we won't be looking to switch unless buyer demands...demand it. I could see this as a cost effective way to offer a lower tier system. I for sure am not saying AMD will not perform as well or better. I guess I mean our customers have faith in Intel and that goes a long way.

 

[PeterScott]

I'll bet MacOS has been running on ARM for a long time already.

Yeah, Apple is bullish on it too; It's called iOS.

It's much more likely that Apple will continue to push that and deprecate MacOS on the lower end mobile segment.

Look at all the new features in iOS 11 and the heavy advertising for iPad.

You can see a clear trend where this is going, and it isn't creating another flavor of MacOS for ARM.

 

[Edrick]

Yeah, Apple is bullish on it too; It's called iOS.

It's much more likely that Apple will continue to push that and deprecate MacOS on the lower end mobile segment.

Look at all the new features in iOS 11 and the heavy advertising for iPad.

You can see a clear trend where this is going, and it isn't creating another flavor of MacOS for ARM.

My guess is that you were not around back in 2003-2005 when the rumors first started about Apple moving to Intel. How many people said it would never happen. Yet here we are.

 

[Sweepr]

If someone leaked the specs of the (remaining) Coffee Lake-S family tonight, what do you guys expect from the Core i5 (65W and 95W) models and the locked Core i7 (65W)? Let's see who's closer to the truth.