Discussion Quo vadis Apple Macs - Intel, AMD and/or ARM CPUs? ARM it is!

[moinmoin]

Due to popular demand I thought somebody should start a proper thread on this pervasive topic. So why not do it myself?

For nearly a decade now Apple has treated their line of Mac laptops, AIOs and Pro workstations more of a stepchild. Their iOS line of products have surpassed it in market size and profit. Their dedicated Mac hardware group was dissolved. Hardware and software updates has been lackluster.

But for Intel Apple clearly is still a major customer, still offering custom chips not to be had outside of Apple products. Clearly Intel is eager to at all costs keep Apple as a major showcase customer.

On the high end of performance Apple's few efforts to create technological impressive products using Intel parts increasingly fall flat. The 3rd gen of MacPros going up to 28 cores could have wowed the audience in earlier years, but when launched in 2019 it already faced 32 core Threadripper/Epyc parts, with 64 core updates of them already on the horizon. A similar fate appears to be coming for the laptops as well, with Ryzen Mobile 4000 besting comparable Intel solutions across the board, with run of the mill OEMs bound to surpass Apple products in battery life. A switch to AMD shouldn't even be a big step considering Apple already has a close work relationship with them, sourcing custom GPUs from them like they do with CPUs from Intel.

On the low end Apple is pushing iPadOS into becoming a workable mutitasking system, with decent keyboard and, most recently, mouse support. Considering the much bigger audience familiar with the iOS mobile interface and App Store, it may make sense to eventually offer a laptop form factor using the already tweaked iPadOS.

By the look of all things Apple Mac products are due to continue stagnating. But just like for Intel, the status quo for Mac products feels increasingly untenable.

 

[jpiniero]

Obviously they won't switch to Threadripper for MP.

If anything Apple will switch to EPYC CPUs. Threadripper's can only be reserved for iMac Pro's.

Workstation apps need lots of mhz. Threadripper would be the appropriate product for Mac Pros.

 

[mikegg]

The funny thing about that initiative, that is supposed to run ARM apps on Mac is that they are supposed to run iOS(!) Application on MacOS, not port MacOS applications to ARM.


Lets face it. If Apple was going to move Macs to ARM, full-stop, they would launch initiative that would port current x86 apps to ARM, not port iOS Apps to x86(!).

They’re going to do both. They will convert their own apps such as Safari to ARM first. They will make partners like Adobe do the same. The rest will follow if they wish to stay relevant in MacOS.

Macs will also gain access to every IPad app.

Apple is going ARM to unify their apps. It’s going to happen.

 

[mikegg]

Workstation apps need lots of mhz. Threadripper would be the appropriate product for Mac Pros.

Threadrippers have a 256gb ram limit which is significantly less than the 1.5th the current Mac Pros offer.

Epyc makes more sense unless AMD creates a custom high frequency parts just for Apple.

 

[Ajay]

They’re going to do both. They will convert their own apps such as Safari to ARM first. They will make partners like Adobe do the same. The rest will follow if they wish to stay relevant in MacOS.

Macs will also gain access to every IPad app.

Apple is going ARM to unify their apps. It’s going to happen.

They can’t make their partners do anything. Apple would need to make a sales pitch to Adobe and others. The two things that will matter most are sales and profits. The very same things Apple will consider when deciding to make a switch. The switch won’t be made on technical merit. Technical merit merely means that an ARM option is feasible. Bottom line is that this is a complex change for Apple and it supplies and vendors. If they hadn’t made such a change before, I would say the chance is zero (Cook is no Steve Jobs) - but they have been down this road before, so there is a chance. Apple knows the cost of such a transition - and, hence, already has a good sense of w or not it is worth it to them and ISVs as well.

I doubt Apple will bring in an outside ARM chip vendor - if they go this way, they will take their time and build a strong offering in house. It would be difficult to build a chip design that could scale from MacBook Pro and iMacs up to Mac Pros (chiplets maybe?).

 

[uzzi38]

Threadrippers have a 256gb ram limit which is significantly less than the 1.5th the current Mac Pros offer.

Epyc makes more sense unless AMD creates a custom high frequency parts just for Apple.

Why make custom ones? High frequency skus are already a thing.

 

[tamz_msc]

Why make custom ones? High frequency skus are already a thing.

3.3GHz isn't high enough.

 

[ksec]

You have to be kidding. Intel has nothing competitive right now. (except the top of the line gaming cpu) Desktop, HEDT, now laptop, and server are all AMD.

Competitive in terms of ARM vs Intel. That post didn't take into account of AMD. On the surface it seems Apple is either going to be Intel x86 or ARM. But again Intel could be competitive with pricing even if it was against AMD.

And it needs to be clear how Thunderbolt 4 / USB 4 is working out. That TB3 controller is still $20 a piece and sold by Intel only.

 

[mikegg]

They can’t make their partners do anything. Apple would need to make a sales pitch to Adobe and others. The two things that will matter most are sales and profits. The very same things Apple will consider when deciding to make a switch. The switch won’t be made on technical merit. Technical merit merely means that an ARM option is feasible. Bottom line is that this is a complex change for Apple and it supplies and vendors. If they hadn’t made such a change before, I would say the chance is zero (Cook is no Steve Jobs) - but they have been down this road before, so there is a chance. Apple knows the cost of such a transition - and, hence, already has a good sense of w or not it is worth it to them and ISVs as well.

I doubt Apple will bring in an outside ARM chip vendor - if they go this way, they will take their time and build a strong offering in house. It would be difficult to build a chip design that could scale from MacBook Pro and iMacs up to Mac Pros (chiplets maybe?).

1. Adobe already has virtually all of its Creative Suite working on iOS and iPad.
2. Adobe is also bringing their entire Creative Suite to work on Windows ARM (source)
3. Apple will provide an x86 emulator just like they did previously when they migrated from PowerPC
4. Apple is going to switch to ARM. They'll start in 2021. Their SDK already makes it easy to build apps for iOS, iPad, and MacOS together.

Whether Apple will move to ARM is no longer debated. They will. It's not a discussion.

Now we're discussing whether Apple should switch to AMD in the next year or so and for which computers.

 

[uzzi38]

3.3GHz isn't high enough.

7H12 ain't a high frequency sku. It's a HPC sku.

The high frequency stuff isn't mass market yet I don't think, you gotta request for it. But afaik they all clock around 4GHz.

That being said, they still don't fit Apple all that well, given the focus on per-core performance I think they also only go up to 32 cores for extra L3 per core.

Well, cloud guys have their own needs and all that.

 

[DrMrLordX]

I think they also only go up to 32 cores for extra L3 per core.

That would get around the alleged kernel panic problem with MacOS.

 

[Gideon]

Threadrippers have a 256gb ram limit which is significantly less than the 1.5th the current Mac Pros offer.

Epyc makes more sense unless AMD creates a custom high frequency parts just for Apple.

That's an artificial limitation they can overcome (for instance with a custom chipset for mac, e g. WRX40. IMO a custom threadripper makes much more sense than custom epyc, IF apple is planning to use AMD (which IMO still seems unlikely).

 

[uzzi38]

That would get around the alleged kernel panic problem with MacOS.

One update; I'm not sure about 32 cores, whether or not it actually hit market (I could have sworn I saw a task manager sceenshot, but oh well), but most the specs of the 8 and 16 cores are already rumoured.

https://twitter.com/x/status/1224707819229843458

Even if a 24 and 32 core one didn't exist though, it should be fairly clear now AMD wouldn't be averse to providing something like this if requested by Apple, so I wouldn't rule out the possibility.

 

[DrMrLordX]

Why would a 16C chip have a 240W TDP at only 3.5 GHz?

 

[mikegg]

That's an artificial limitation they can overcome (for instance with a custom chipset for mac, e g. WRX40. IMO a custom threadripper makes much more sense than custom epyc, IF apple is planning to use AMD (which IMO still seems unlikely).

Yes, it's a limit to differentiate Threadripper and Epyc. When I said "custom part", I just meant different bins with different features enabled/disabled. I didn't think AMD would have to create a different architecture for Apple.

Apple, if they choose to use AMD, will get the hardware configured however they want at the best possible price.

 

[uzzi38]

Why would a 16C chip have a 240W TDP at only 3.5 GHz?

We don't know if that's base or boost. We also don't know in what workload that frequency is rated for if it's boost.

No, I'm dumb, it is a base clock, not a boost clock.

 

[DrMrLordX]

We don't know if that's base or boost. We also don't know in what workload that frequency is rated for if it's boost.

No, I'm dumb, it is a base clock, not a boost clock.

Must have one hell of a boost clock then.

 

[moinmoin]

Why would a 16C chip have a 240W TDP at only 3.5 GHz?

Look at the L3$ size. It's an insane package with 8 CCDs where only one core is activated in every CCX.

 

[uzzi38]

Look at the L3$ size. It's an insane package with 8 CCDs where only one core is activated in every CCX.

The epitome of per-core performance :^P

Stupid clocks and stupid cache per core.

 

[Glo.]

They’re going to do both. They will convert their own apps such as Safari to ARM first. They will make partners like Adobe do the same. The rest will follow if they wish to stay relevant in MacOS.

Macs will also gain access to every IPad app.

Apple is going ARM to unify their apps. It’s going to happen.

Safari already is ported to ARM, whats more, iPads already use DESKTOP version of Safari.

If you are so sure, about the reality that Apple will switch to ARM, I will ask you a simple question.

When can we expect that Apple will migrate their own largest, and most important creative applications to ARM: Final Cut Pro X and Logic Pro?

 

[Glo.]

3. Apple will provide an x86 emulator just like they did previously when they migrated from PowerPC
4. Apple is going to switch to ARM. They'll start in 2021. Their SDK already makes it easy to build apps for iOS, iPad, and MacOS together.

Nr 3 is only your speculation. Not a fact, unless you have insider information.

Apple moving to ARM Macs will start happening the moment when they will port Final Cut Pro X and Logic Pro to iPads. Not before. So far it hasn't happened, and yet - everybody says that it is a done deal that Apple is moving to ARM on Macs.

 

[DrMrLordX]

Look at the L3$ size. It's an insane package with 8 CCDs where only one core is activated in every CCX.

Ohhh I get it. That thing could potentially boost to 4.5 GHz on every core.

 

[Ajay]

1. Adobe already has virtually all of its Creative Suite working on iOS and iPad.
2. Adobe is also bringing their entire Creative Suite to work on Windows ARM (source)
3. Apple will provide an x86 emulator just like they did previously when they migrated from PowerPC
4. Apple is going to switch to ARM. They'll start in 2021. Their SDK already makes it easy to build apps for iOS, iPad, and MacOS together.

Whether Apple will move to ARM is no longer debated. They will. It's not a discussion.

Now we're discussing whether Apple should switch to AMD in the next year or so and for which computers.

Apparently you didn't read the thread title. We are, in fact, discussing exactly what I posted about.
Glo. already pointed out the fallacy of No. 3.
As to No. 4

Kuo believes that Arm-based processors will significantly enhance the competitive advantage of the Mac lineup, allow Apple to refresh its Mac models without relying on Intel's processor roadmap, reduce processor costs by 40 to 60 percent, and provide Macs with more hardware differentiation from Windows PCs.

Kuo 'believes' - it doesn't say that Kuo 'knows'? It's a research note and, almost certainly, has a caveat that this note is speculative and bears certain risks. Maybe he's right, maybe not. We do not have any concrete information at this time.
We don't even have leaks of engineering prototypes - nothing concrete at all, AFAIK.

If you have any proof of your claims, please post them. I'm quite fine with being proved wrong on the basis of facts.

 

[Ajay]

Workstation apps need lots of mhz. Threadripper would be the appropriate product for Mac Pros.

They need allot of computer performance, doesn't matter how they get there. Obviously though, some sort of reasonable clock rate will be needed (3GHz? maybe, we don't really know).

 

[mikegg]

Safari already is ported to ARM, whats more, iPads already use DESKTOP version of Safari.

If you are so sure, about the reality that Apple will switch to ARM, I will ask you a simple question.

When can we expect that Apple will migrate their own largest, and most important creative applications to ARM: Final Cut Pro X and Logic Pro?

I meant desktop Safari which is obviously still x86. It shouldn't be too hard to port it over to ARM given that it's been ARM-compatible since the first iPhone.

I'm sure Apple will switch to ARM. It's no longer debatable. Having apps work seamlessly between iOS, iPadOS, and MacOS is far more important to Apple than not pissing off a portion of their Mac userbase.

But how would I know exactly when they'll port Final Cut Prox and Logic Pro to ARM? This question doesn't seem relevant. But if I have to guess, Apple will probably release it at the same time Macbook Pro switches to ARM. Adobe Premier already works on iPad which is a good sign that Apple is fully capable and willing to port its own video editing software over.

 

[Glo.]

The biggest logical fallacy that people genuinely believe, that is true, about Apple porting Macs to ARM is very simple.

They are making ARM Software run on x86 architecture, not the other way around. If they would be moving Macs from x86 to ARM shouldn't they do exactly the latter: already announce option for porting x86 Apps to ARM?

ARM Macs are not going to be the same as MacBook Air/Pro, and iMac Pro/Mac Pro.

ARM Macs will be like Apple TV, that has Screen, keyboard, and access to all of Apple most important services: Apple Arcade, Apple Music, Apple TV. There was a rumor some time ago that Apple would release a Mac targeted towards gamers.

What is the biggest Gaming platform in Apple ecosystem? Yeah... I can easily see AppleTV-like box that can easily work both like Mac or... Gaming Console.

I meant desktop Safari which is obviously still x86. It shouldn't be too hard to port it over to ARM given that it's been ARM-compatible since the first iPhone.

I'm sure Apple will switch to ARM. It's no longer debatable. Having apps work seamlessly between iOS, iPadOS, and MacOS is far more important to Apple than not pissing off a portion of their Mac userbase.

But how would I know exactly when they'll port Final Cut Prox and Logic Pro to ARM? Doesn't seem relevant.

It is 100% relevant. Because if Apple will not port FCPX and Logic Pro to iPads - they are NOT MOVING THEIR MACS TO ARM! Its that simple. That is not how Apple works, and how Apple does things. Apple's third biggest revenue stream is rapidly becoming Services. Logic Pro is used by hundreds of thousands of musicians which are posting their music in Apple Services. Apple designed Mac Pro, and FCPX specifically for people that are in Apple ecosystem for Video Production, mainly because they want people that use AppleTV Services to also produce their films on Macs.

Apple hasn't ported both of those Apps for simple reason. There is not enough performance in ARM CPUs. Remember - porting all of their platforms to ARM would be justifiable only if it would give performance benefits for people who need this, not just toys on which they can watch Youtube.