Intel Says 10nm Chip Development is On Track

[oldMac]

How can one compare ARM to x86-64 CPUs?

These are in totally different league.

They really are not that far apart at this point. Apple's ARM designs have made huge gains with each annual generation and Intel's processors aren't progressing at anywhere near the same pace. It's quite possible that Apple will hit a wall around the same place that Intel is now, but given the pace, we (and Apple) will know that very soon if they don't know in the lab already.

On top of getting close to performance parity, the real story is how much more energy efficient Apple's processors are compared to Intel at the same performance levels. Even without hitting performance parity with Intel's fastest chips, it's likely that Apple could deliver a decently performing laptop *now* with much better battery life than a similar Intel-based laptop.

 

[bdkennedy1]

Only 6 years to go until Apple even considers putting them in anything then.

Based on your opinion. Apple under clocks their processors. No one except Apple knows what they are capable of.

 

[cmaier]

But will the software that people NEED to use (not Twitter or fart apps) actually work on ARM? Windows on ARM doesn't seem to be succeeding in that regard.

Apple has much more experience with successful architecture shifts. They’ve also built the infrastructure (mac App Store, bitcode, etc.) to make it work.

 

[oldMac]

Lol people are scoring almost 10k single core and over 50k multi core on 14nm i9 9900k
View attachment 797587 Intel is still 5 years ahead of anyone and will be on another level when they get 10nm done. You people know that apples 7nm is not even close to being a true 7nm process.

Here for you people that are obsessed with geekbench scores.

Please somebody please name one freaking CPU that can beat an Intel 14nm CPU that is going on 3 years now.

Just pointing out that this is either fake or overclocked (i.e. not production-reliable hardware.) Reference: https://browser.geekbench.com/v4/cpu/9742328

Real numbers are closer to 6250/33037. Still impressive, but Apple has a friggin' phone scoring 4795/11218. If Apple really put their efforts into developing a Desktop processor on ARM (no doubt they are), they could likely do better than their ultra-low-power phone processor scores.

My prediction is that will be *painfully* obvious to Intel (and everyone else) by the end of 2019.

 

[cube]

you do realise that those numbers are pretty much just names, right? intel's 10nm (if it works) is better than TSMC's 7nm. so TSMC 5nm is probably on par with intel's 10.

No, if the trend continues, TSMC 5nm will be like Intel 7nm (and it could be that TSMC 3nm will be like Intel 7nm).

 

[Unregistered 4U]

Based on your opinion. Apple under clocks their processors. No one except Apple knows what they are capable of.

Seeing as how they are designed to hit just over 2 GHz, one must assume that some chips on that wafer could possibly clock far higher given the right voltage and cooling.

And, I wouldn’t doubt that a few lucky ones make their way back to Apple’s research labs

 

[cube]

AMD's current CPUs are a mix of 12nm and 14nm and expected to be on 7nm in 2019.

With the current processes less advanced than Intel 14nm and the next one comparable to Intel 10nm.

 

[Unregistered 4U]

But will the software that people NEED to use (not Twitter or fart apps) actually work on ARM? Windows on ARM doesn't seem to be succeeding in that regard.

I would think that’s primarily because Windows has a LOT of choice. As a developer, do I want to develop for the leading edge making a tiny amount of money, or the HUGE middle and make a lot of money? With Apple, the proposition will be more like “Do I still want to be a macOS developer?” Because if Apple phases out Intel, the developer doesn’t have a choice. They either support ARM or fold. Sure, they could continue trying to support the old architecture, but all someone has to do is replicate their app in ARM and their customers would dry up overnight.

 

[Fmello]

If Apple did design their own chips for their Macs, what would you hope to see in that chip if it came to fruition?

 

[Mr. Dee]

Still on my 14 NM Broadwell, but I won't be considering a MacBook Pro upgrade until Apple actually does get 10 NM chips from Intel. Although the optimizations Intel has added to minor revisions like Skylake, Kabylake and Coffeelake are welcome, its still just a rebadging.

On top of that, honestly would be wary about jumping on 10 NM early, which puts my upgrade strategy around late 2020 on track.

I am starting to feel the growing pains with my Early 2015 MBP though. Basic apps like Office 365 and single Windows 10 VM I use are struggling.

 

[Unregistered 4U]

They really are not that far apart at this point. Apple's ARM designs have made huge gains with each annual generation and Intel's processors aren't progressing at anywhere near the same pace.

And, whatever Apple comes out with won’t HAVE to perform better than Intel at all tasks, it ONLY has to run macOS faster than Intel, Logic faster than Intel, Final Cut Pro faster than Intel, Pages faster than... you get the picture. If they can show macOS on ARM processing a Pixelmator image faster than macOS on Intel, they’ll have some day 1 converts.

Speed may matter more in emulation, but emulation is not a big part of Apple’s future. Try going to to Apple.com and searching for Emulation or Virtualization. Unless you already knew about Boot Camp, you would have found nothing that points to, say, being able to run other OS’.

 

[iPadified]

Intel is not gonna cry by loosing Apple, is a very small customer for them...as a former Intel eng. I know the hardship of building x64 architecture chips on anything smaller than 12 nm. We are talking about playing with single atoms almost...
[doublepost=1540230559][/doublepost]

arm architecture, simpler than x64. That kind of architecture Intel is able to fit in 5 nm long time ago...It's not the same to build a car than a big commercial plane.

Care to elaborate? Why would x64 be more difficult than ARM? 10 nm defines the feature size such as a gate or part of a gate. In my ignorance I though a gate was a gate and the configuration of transistors and gates defines if is a x64 or ARM architecture.

1D size measures in processor industry is mostly marketing... Transistor per square cm would be a better measurement.

 

[manu chao]

Glad they are working on getting 2 year old tech to market.

Meanwhile, TSMC is working on 5nm.

Intel is a bit more conservative with its process size labelling. Intel's 10 nm is closer to TSCM's 7 nm than to TSCM's 10 nm process.

 

[Oculus Mentis]

With custom-designed Mac chips, Apple will no longer be forced to delay updates due to Intel's manufacturing issues

Would really owning your own design give you total control over the uncertainties of the manufacturing process?

 

[manu chao]

Show me.

While not a perfect match, Intel's 10 nm process is grouped under the "7 nm" processes label:

https://www.semiwiki.com/forum/content/7544-7nm-5nm-3nm-logic-current-projected-processes.html
[doublepost=1540278484][/doublepost]

Total B.S. - There have been multiple chip iterations. Apple could have been updating each time to a new chip. They also could have done subtle updates, adding Thunderbolt 3, new GPUS, etc. I understand they built themselves into a corner, but didn't they realize that 2-3 years ago?

They did, that's when they decided to kill off the Mac Pro and started work on the iMac Pro that shipped almost a year ago. It was only later when they had a change of hard and moved to resurrect the Mac Pro.

 

[BeamWalker]

With custom-designed Mac chips, Apple will no longer be forced to delay updates due to Intel's manufacturing issues, and custom chips will give the company more control over design, better profits, and a way to differentiate its products from competing PCs.

Because No-Competition always works out great for the customer.

 

[opeter]

Because No-Competition always works out great for the customer.

There is competition in form of Windows, Linux etc., not to mention the many other manufacturers of hardware.
Apple is a long time not the only option.

 

[manu chao]

Companies who use Intel processors and who don't make yearly computer updates:
Apple

There hasn't been a single year where Apple hasn't updated the Macbook Pro plus another laptop line (MB: 2006-2010, 2015+, MBA: 2008-2015). The iMac has also been updated every single year except 2016.

Apple simply has two computer categories: Those that get updated yearly and those that don't. The former include the MBP, MB, MBA until 2015, and the iMac (with one exception and a few asterisks). The latter include the Mac Pro, Mac Mini and their cheapest laptop model (13" non-retina MBP from 2013 on, 13" MBA from 2016 on).

Whenever somebody mentions the first category, others chime in with "But but, what about the second category?", and vice versa.

 

[Michael Scrip]

There hasn't been a single year where Apple hasn't updated the Macbook Pro plus another laptop line (MB: 2006-2010, 2015+, MBA: 2008-2015). The iMac has also been updated every single year except 2016.

Apple simply has two computer categories: Those that get updated yearly and those that don't. The former include the MBP, MB, MBA until 2015, and the iMac (with one exception and a few asterisks). The latter include the Mac Pro, Mac Mini and their cheapest laptop model (13" non-retina MBP from 2013 on, 13" MBA from 2016 on).

Whenever somebody mentions the first category, others chime in with "But but, what about the second category?", and vice versa.

My point was... whenever there are certain Macs don't get updated... it's not because of Intel. My reply was to a comment about "Intel delays"

You're right about the first category. Let's call those the "good" Macs. Those get updated regularly.

But when you only have 7 product lines... and 3 of those 7 sit abandoned... it sure looks funny.

Again... it's not Intel's fault that the Mac Pro, Mac Mini and Macbook Air haven't been updated in ages.

 

[donileo]

Intel is not gonna cry by loosing Apple, is a very small customer for them...as a former Intel eng. I know the hardship of building x64 architecture chips on anything smaller than 12 nm. We are talking about playing with single atoms almost...
[doublepost=1540230559][/doublepost]

arm architecture, simpler than x64. That kind of architecture Intel is able to fit in 5 nm long time ago...It's not the same to build a car than a big commercial plane.

Given you are saying that the complexity of the chips is different which I assumed, can you explain why “arm” chips are being touted as desktop class and desktop replacements? If they are so simple why are they being said to replace intel in Macbook Pros in the future. In the server space they are already making headway with 64 core chips etc etc.

 

[asiga]

Apple is not doing it better with their T2 chips (2018 MBP BridgeOS KPs not fixed yet, even after countless supposedly fixes that didn't fix it). I prefer Intel, but that's irrelevant because I won't be a Mac user anymore when MacOS becomes as user-unfriendly and user-controlling as iOS.

 

[alexmarchuk]

Given you are saying that the complexity of the chips is different which I assumed, can you explain why “arm” chips are being touted as desktop class and desktop replacements? If they are so simple why are they being said to replace intel in Macbook Pros in the future. In the server space they are already making headway with 64 core chips etc etc.

ARM chips are definitely not in the same class as desktop chips. My 8700k runs circles around them in complex functions.

 

[manu chao]

My point was... whenever there are certain Macs don't get updated... it's not because of Intel. My reply was to a comment about "Intel delays"

You're right about the first category. Let's call those the "good" Macs. Those get updated regularly.

But delays at Intel have delayed updates for those 'good' Macs too (by months, not years). And talk about MBPs (probably the most regularly updated Macs) being late has been quite common over the last couple of years.

Thus when somebody says Intel is the reason for some delays in new Mac releases, part of the readership thinks of those situations when the 'good' Macs were late due to Intel's delays and another part of the readership thinks of those Macs were Intel was not at fault. The very same sentence is read differently depending on what you think the author tried to convey. If you think Macrumors are Apple apologists, then any sentence with the words 'delays', 'Intel', and 'Macs' might seem as an attempt to whitewash Apple. But if for some reason you are a bit more prone to give people the benefit of the doubt, you just assume that the most innocent explanation for such a phrase is the correct one.

But when you only have 7 product lines... and 3 of those 7 sit abandoned... it sure looks funny.

Before the MBA was 'abandoned' (which happened essentially in 2016 when it didn't get an update alongside the other laptops), the non-abandoned products (ie, all except Mac Pro & Mini) probably comprised 95% of all Mac sales.

Moreover, your contrasting of how 'normal' manufacturers update their products and how Apple does it is a bit facetious. Apple simply has far, far fewer models and thus releases updates only a few times per year. Often probably only two or three times a year.

Again... it's not Intel's fault that the Mac Pro, Mac Mini and Macbook Air haven't been updated in ages.

And I guess my point is that nobody ever claimed such or even tried to imply such a thing. You are seeing something that I don't see.

 

[Val-kyrie]

Hardware support lasted until being declared "obsolete" in 2013, 7 years after the last PPC machine. I see that as an emminently reasonable amount of time.

Hardware support is not the issue. How many more OS updates did PowerPC Macs receive?

 

[cube]

Hardware support lasted until being declared "obsolete" in 2013, 7 years after the last PPC machine. I see that as an emminently reasonable amount of time.

You could install Windows 10 on a PC from 2006.

You could install a latest Linux on a PC from 1989.