Intel Restates Support for Apple During Transition to Apple Silicon

[1144557]

Chips today are made up from building blocks. You don't design monster chips so much as connect them together like lego. Fab a test run can be real quick nowadays.

Agreed. And people think they are just going to stick an A12z (whatever iOS device chip is that year) into the Macs and boom done!

They never said they are re-using that year's iOS device line of chips; that is only the development platform right now. They probably have a ton of those chips sitting around to stick in Mac Minis as a test bed (especially with slow sales due to COVID). It's convenient and they already have a bin of those ready. They didnt lay out their strategy so that is all assumptions.

They may not HAVE a Mac Pro level (Xeon/Threadripper type) competitor chip ready, that is why it will be years of phasing still using Intel.

And likely why they are starting with the consumer grade stuff first; the general i-5/7/9 which should be relatively easy top better performance and thermally with their silicon

They have some pretty dang good chip people. Everyone said this same line when Apple started making their A-series, oh it wont compete etc, and they blow the competitors out every single year.

 

[zeroID]

Actually no special need for a Mac just the iPad battery life is too short and eventual replacement is not possible at all on iPad Pro 10.5 !
Why Apple is not redesigning the iPhone and the iPad to make the battery swap possible at user level ???

 

[dogslobber]

Agreed. And people think they are just going to stick an A12z (whatever iOS device chip is that year) into the Macs and boom done!

They never said they are re-using that year's iOS device line of chips; that is only the development platform right now.

They didnt lay out their strategy so that is all assumptions.

Yep. The chips will likely be highly threaded (many cores like dozens) which is why there's the heavy emphasis on Grand Dispatch in the engineering sessions. But these chips will blow our minds when they're benchmarked.

 

[ACE76]

ummm, its a multi-core chip - dude, and runs a multi-threaded variant of Unix (has for many years). I'm not sure where you are going with this comment, but it sounds wrong. Sure the A12z runs squarely in the 4-core i5-i7 performance range, but I'm pretty sure the released Apple silicon will be more like an A14x++, maybe have more cores and extra GPU oomph seeing it is not for a mobile application, but for a laptop.

And the thermal issues associated with iPhone/iPad implementations will not be the same on a laptop.

Finally, Apple licenses the ARM IP, and custom designs it on chips, so there is a lot more than 1 company behind this. And you know ARM, the smarts in Snapdragon, Exynos, and many others.

So yah, not as negative as you make it sound.

On the upper end of workstation chips, yah, that might be awhile, but the technology is scalable so we can wait and see

If ARM was so powerful, you would have seen an entire market for it in the desktop/laptop market. Microsoft also released a few products running on ARM and they were abysmal. If all you need is a machine for browsing and playing smartphone level games, sure, go for it. I would never pay the prices Apple charges for computers to get a ARM based machine.

 

[cardfan]

To be fair you can get an idea of the A12Z's performance by looking at the iPad Pro 2020 lineup though I guess it might be capable of more performance on an actively cooled Mac.

which isn’t anything to brag about. An iPad runs a mobile os with limited apps.

 

[elvisimprsntr]

At the time of the Intel transition I remember seeing the word "Mactel", so is this the day of "Macarm"?

"ARMac" is better.

Jony Ive gave us "Mac-n-Cheese"

 

[AussieStu]

I would have liked AMD options in the last few years. Choose better range in CPUs for price.
Doesnt give much incentive to buy a new Mac in the next 12 months now.

 

[johnnytravels]

I am astounded by the way most people seem dead certain that 1.) Apple has something tremendous up their sleeves which they just haven’t announced yet and 2.) will stick to the timeline suggested.
This is Airpower and the timeline following the first Mac Pro announcement all over again.

Don’t get me wrong, I would love for Macs to be more powerful and efficient at the same time and don’t care much for Intel. At the same time, all I trust Apple to reliably control is the release of their annual iPhone iteration. The rest is subject to how well they can integrate Cook’s penny pinching with actual technological development.

 

[cmaier]

Is the industry moving away from x86? Or is Apple just an isolated case?

I haven't heard about Dell, HP, Lenovo, Asus, Acer and every other PC manufacturer exploring options other than x86.

Those companies may have some models running Qualcomm chips, for instance, but those seem to be experiments at best. None of those sound like an edict to abandon x86.

The server market is slowly moving to Arm as well, and for good reason. Power/watt is incredibly important in large server farms, where if you want more capacity you need to build new power distribution and cooling for entire buildings. Now the desktop has cracked too.

Reminds me of when Linux first appeared. “It’s not real Unix. It’s not compatible with stuff. Blah blah blah. It’s a toy.”

Just you wait.

 

[AussieStu]

which isn’t anything to brag about. An iPad runs a mobile os with limited apps.

Thats what I was thinking. That hardware running a Full OS like on a Mac could see average increase. Prob just better to do a deal with AMD for same increase in speed over Intel without all the extra coding and dev crap to deal with.

 

[cmaier]

I am astounded by the way most people seem dead certain that 1.) Apple has something tremendous up their sleeves which they just haven’t announced yet and 2.) will stick to the timeline suggested.
This is Airpower and the timeline following the first Mac Pro announcement all over again.

Don’t get me wrong, I would love for Macs to be more powerful and efficient at the same time and don’t care much for Intel. At the same time, all I trust Apple to reliably control is the release of their annual iPhone iteration. The rest is subject to how well they can integrate Cook’s penny pinching with actual technological development.

They have YEARS of track record of silicon success, and are really the only computer maker to have successfully made 2 and a half such transitions successfully already. There’s good reason to trust them.

 

[ctrlzone]

reliably control is the release of their annual iPhone iteration

with Apple no longer relying on Intel, we can assume the same for the Mac lineup.
"frequent" iterations on a regular basis.

 

[johnnytravels]

Yes, an unreleased product sure says a lot.

They have YEARS of track record of silicon success, and are really the only computer maker to have successfully made 2 and a half such transitions successfully already. There’s good reason to trust them.

yet they couldn’t put together a desktop computer from essentially off the shelf compinents.

I really hope you are right though!

 

[justin216]

If ARM was so powerful, you would have seen an entire market for it in the desktop/laptop market. Microsoft also released a few products running on ARM and they were abysmal. If all you need is a machine for browsing and playing smartphone level games, sure, go for it. I would never pay the prices Apple charges for computers to get a ARM based machine.

Software lock-in is a thing. Microsoft has tried a few times to push Windows on ARM, including most recently with the Surface Pro X, which they called a halo device to try to push the paradigm and a decently specced unit is just over $1000 without accessories -- add the accessories, and you're at $1500, for a WoA device.

Microsoft has the burden of more legacy software by a much larger number of users, coupled with drivers for a larger number of devices, OEMs with different implementations, etc. They, quite frankly, have a dramatically harder time shifting architectures. The SPX is quite fast when you're using ARM64 binaries (recently tried VSCode insider build for ARM64 with a bunch of extensions and it was lightning fast), but then falls over when you try to use x86 apps on it -- janky, slow, etc. We'll obviously see if Apple's Rosetta 2 performs better for these tasks, it seems like it will, but time will tell.

This will be the third time Apple has shifted architectures, their more closed ecosystem lends itself to being able to make the switch more manageable.

There are also more and more servers being deployed with ARM chips, granted the workload on a server is very specific, controlled, and can be tailored to the business objective.

 

[AndyDiamond]

This is going to be very cool!

 

[cmaier]

yet they couldn’t put together a desktop computer from essentially off the shelf compinents.

I really hope you are right though!

It’s not that they COULDN’T. It’s that they didn’t want to.

 

[scupking]

The new 5nm A14 and A14x are going to be amazing chips when we see them at the end of the year. I can only imagine what a desktop and laptop version of that chip will be like!

 

[theluggage]

The damage to Intel is not necessarily the loss of Apple's immediate business. But the long-term, cumulative effect of a growing movement away from x86, and the industry's reliance on Intel as a whole.

This.

Apple may only have a fraction of combined PC sales, but the Macs are probably the highest-profile individual models and are hugely influential, not to mention the leverage Apple get from the iPhone. When did you last see a Dell, Lenovo or even Microsoft (hardware) product launch get the same sort of press coverage (including mainstream sites like the BBC) that Apple does? Although Apple have been slipping a bit in recent years, ever since they launched the Powerbook 100, the entire PC laptop industry has been following Apple's lead (and, yeah, without the iPhone, Android would have been a keyboard/jogwheel-driven Blackberry knock-off).

All efforts so far to make an ARM-based personal computer - the MS Surface X, even ARM-based Chromebooks - have been a bit half-baked and suffered from the software deadlock: developers won't support machines that nobody buys, people won't buy machines without software support.

Apple are in a position to force the issue. Now that Tim has stood up and said that Macs will switch to ARM over the next 2 years then you can be fairly certain that it will happen and that a critical mass of native software will appear.

If the ARM Macs are successful - even on the scale of current Mac sales - they are going to do a lot for the credibility of ARM as a personal computer/workstation processor - and others may follow.

Don't forget, all you "never gonna happen" folks, the mobile sector has already decimated (at a minimum in the "reduced by 10%" sense of the term) the PC market and there, Microsoft and Intel lost out dismally to ARM and Unix-like-OSs. The "Wintel" monopoly was founded on the idea that binary compatibility was king - but technology has moved on from that now (heck, it was ready to move on in 1981, just before flippin' IBM stuck their name on a me-too CP/M knockoff running on a kludgey stopgap pseudo-16-bit processor and turned the industry to stone for 30 years). All it really needs is a change in mindset...

Still, Intel could always make their own ARM-based CPUs. Actually they did used to make an ARM-based CPU (they inherited the StrongARM from Digital/Compaq) so it wouldn't be a first.

Lot of furious people who just bought the new MBA, I’d imagine.

Yes - you buy a new computer and suddenly find out that, in 6-12 months' time, something newer and better will be available. Oh the humanity...

Still, if you're running high-end x86 software under Windows bootcamp then now is the time to buy a new Mac. Personally, if I wanted to do that, I'd buy a PC...

 

[SuperCachetes]

"Apple is a customer across several areas of business and we will continue to support them*" said an Intel spokesperson.

"*on our own timetable..."

To me, unhitching from the bumpy Intel roadmap is huge. Being able to align the timetable with product releases and the ability to tailor a SoC to an exact device (as opposed to having to pick from another company's catalog) is next-level. Everyone is looking at software compatibility, but I really think the design and manufacturing control are the game-changers here.

Correct... but if PC manufactures start using AMD instead of Intel... that's still x86

The topic was about the industry leaving x86

Exactly. Apple is a singular entity that can control its own destiny in a lot of ways. Somehow getting Microsoft to 1) want to turn over its apple cart (no pun intended) and move to ARM/RISC, and 2) having them convince all the OEMs betrothed to Windows to move along with them in a reasonably simultaneous manner is just not going to happen.

ARM CPUs will never be as powerful as x86. That's just a plain fact and always will be.

Just a plain question: are you an Intel employee, a bot on the fritz, or just a comedian?

 

[Brian Y]

Even though Apple has a tendency to buy only higher-end chips from Intel... I don't think this will affect Intel too much.

There have been theories that suggest the opposite. Many people have suggested that Apple actually buys lower-quality binned chips from Intel. As has been noted by many people swapping chips in iMacs, etc. (e.g. if you upgrade a 2017 iMac to the same i7 7700 that came with it, you'll end up with a slightly faster machine than the stock 7700).

 

[B4U]

I mean.... what else are they going to say publicly?

"We will continue to pay child-support even though we are the best partner that he/she has ever met." 🤣
Does this sound like basically what Intel said in a sense?

 

[ruka.snow]

I would have liked AMD options in the last few years. Choose better range in CPUs for price.
Doesnt give much incentive to buy a new Mac in the next 12 months now.

AMD CPU's would require a lot of developers to at least recompile their code. Apple targets Intel specific features. AMD are also only just recently competitive, how do we or Apple know that by the time they get a new design together that Intel wouldn't have gotten its act together?

 

[1144557]

How many times does it have to be said. Software + hardware. Software + hardware. Software + hardware. We've heard this over and over even in iOS vs Android debates. Controlling both sides you dont need NEAR the amount of resources to get the same experience.

This is not equivalent to Microsoft etc trying it and getting devs etc on board. They can barely get devs on board for basic stuff. EVERYTHING will be reworked for ARM in the OS and apps. Devs wont have a choice here. Anything else is assuming right now.

And no one ever said oh yeh ARM is going to DOUBLE/TRIPLE Intel performance. No one ever said there would be ANY performance increases. What isnt fast enough right now exactly? (which a lot of is thermal issues not raw speed).

But there are no promises here begin made; not that I saw in the presentation. If they perform at least well as Intel right now out of the gate and thermally better that is still a win for Apple (not paying Intel for chips either).

 

[cmaier]

No they don't. ARM CPUs will never be as powerful as x86. That's just a plain fact and always will be. Apple is literally dropping support for all the people who used their hardware for development. Let's see that Ax CPU compete with a high end 16 core/32 thread x86 CPU when it comes to 3D modeling or 8k video editing. These new Macs will be good for browsing and light duty tasks.

Why wouldn’t a 32 core ARM trounce a 16 core/32 thread x86 at a fraction of the power? After all, cores are better than threads, and ARM cores are much smaller than x86 cores, especially when the x86 cores have to support hyperthreading. X86 also has bigger instruction decoders, microcode sequencers and ROMs, bigger load/store units, etc., all of which requires more complicated pipelines, which means you need bigger branch predict units to make up for that, all of which takes more transistors, generating more heat, using more power, and taking more space.

 

[UltimaKilo]

It’s likely going to be a slow process for Apple to build out processors that can rival Intel’s higher end offerings.

A new 2021 MacBook running on a modified, fan cooled A14, could be a good test run. However, I’d be shocked if they have something suitable for the MacBook Pros before 2022 and later for desktops.

The move away from x86, if it must come, should be gradually done over a 3-4 year period.

Next to zero? I think the A4 - A13 speak otherwise. Sure, intel is in the business much longer, but clearly intel is facing issues with their tick-tock-tock-tock cycle and intel doesn't seem to be serious in GPU (while Apple is doubling down on GPU on the Ax chip).

Intel is literally about to get into the GPU space. They’re very serious about it since it’s where the money is.