Bloomberg: Apple's First Arm Mac to Launch by 2021 With 12-Core Processor

[PickUrPoison]

It isn't hate - it's just that the Reality Distortion Field never affected me. I have always known that Steve Jobs was nothing more than a 2nd rate P.T. Barnum. I am technically literate, and the computer is a tool. I came for the powerful hardware (PowerMac G3 then G4, G5, then Mac Pro 1,1, & 4,1) and well designed software. That isn't Apple anymore.

My judgements on Apple products are based on use and in comparison to what is available on the PC side. I am irritated that I stayed with Apple as long as I did - Had Apple actually published a roadmap, I could have (and would have) jumped ship in 2017, along with everybody else.

The Mac Pro series of computers started out as general desktop workstation class computers. With the release of the trashcan, it became a dongle for Final Cut. After the "Mia Culpa" tour of 2017, I (and a lot of other people) were expecting a return to a general purpose workstation.

That isn't the Mac Pro 7,1. It is another dongle for Final Cut. EVERY subsystem in it was obsolete on the day it launched. It's CPUs are a 14nm++++ cul-de-sac that even Intel is dumping. The GPUs are either 2 generations back (580 is a 4 year old card (Polaris) that sells for $150. The Vegas are 1 generation back (AMD has ceased manufacturing Vega 56 GPUs, so the iMac Pro will get an update). The WX5700 (RDNA1) hasn't shown up yet, but RDNA 2 GPUs will ship next quarter. The I/O is PCIe 3.0 - PCIe 4.0 launched last year, PCIe 5.0 launches next year. And then there is the $4,500 case and the $700 wheels.

If Apple had built the Mac Pro 7,1 around Threadripper - I would have happily bought one, because the hardware would once again be cutting edge - and since what I do with my computer is CPU, GPU, and Ram intensive, I wouldn't have had to spend six months on a migration strategy away from Apple.

OTOH, I now have more horsepower for less than a third of the price of a 7,1. Windows 10 is just as bullet proof as OSX. And much more useful, since I don't have to toss all of my 32-bit apps.

If you don’t need or want MacOS, you can probably save money buying a PC. It’s been that way for decades.

PS Threadripper is PC class, and not sufficient or appropriate for the workstation market. EPYC is AMD’s equivalent to Intel’s Xeon. But you’re unlikely to see that on the Mac Pro, because Apple is unlikely to switch to AMD.

 

[turbineseaplane]

@ssgbryan Right on man - I totally feel you

For me, if it weren't for Hackintosh, I wouldn't even be using macOS anymore as I simply refuse to get my wallet shredded apart by Tim for overpriced and underpowered hardware (or some combo thereof) all the time.

I guess it remains to be seen how long using a Hack will be viable, but I expect it will be for at least 3-4 more years.

 

[PickUrPoison]

@ssgbryan Right on man - I totally feel you

For me, if it weren't for Hackintosh, I wouldn't even be using macOS anymore as I simply refuse to get my wallet shredded apart by Tim for overpriced and underpowered hardware (or some combo thereof) all the time.

I guess it remains to be seen how long using a Hack will be viable, but I expect it will be for at least 3-4 more years.

Apple isn’t the right solution for everyone. Every year, only about 200 million are willing (and able) to pay for iPhone/iOS, 40 million for iPad/iPadOS, 30 million for Watch/WatchOS and 20 million for Mac/MacOS. There’s plenty of room for other products, and many, many alternatives. Choice is good 🙂

I’ll never understand why those who don’t like Apple’s products and/or prices are so obsessed with them, but hell I’m all for whatever increases one’s happiness. If that’s complaining endlessly about Apple’s products/prices, that’s cool with me. There are worse hobbies I guess!

 

[Jury]

AMD's roadmap will be ahead of Intel's for at least the next 3 years. By the time Intel gets to 10nm in quantity (equivalent to AMD's 7nm), AMD will be on Zen 5 and 5nm. TSMC is already seeing good results on 3nm.

Better performance per watt? Seriously? - you obviously haven't paid attention over the last 3 years:

1. Intel reports TDP based on core clock speeds - AMD reports TDP based on boost clock speeds.
2. AMD 3950X (16 cores/32 threads) draws 105 watts (at boost). Intel's new 10 core/20 thread i9 draws 125 watts at base - 300 watts at boost (which is why it needs a new mother board to handle the power draw).

So, apparently, you don't actually know.

Agreed, that's why Apple won't be abandoning x86 if the performance per watt is better with AMD and the mac installed base is far higher than PPC era at approximately 5 times more.

 

[ssgbryan]

Apple isn’t the right solution for everyone. Every year, only about 200 million are willing (and able) to pay for iPhone/iOS, 40 million for iPad/iPadOS, 30 million for Watch/WatchOS and 20 million for Mac/MacOS. There’s plenty of room for other products, and many, many alternatives. Choice is good 🙂

I’ll never understand why those who don’t like Apple’s products and/or prices are so obsessed with them, but hell I’m all for whatever increases one’s happiness. If that’s complaining endlessly about Apple’s products/prices, that’s cool with me. There are worse hobbies I guess!

Every year it appears that Apple isn't the solution for more and more people that actually create. Apple has moved from a Hardware/Software solution for creative professionals/amateurs to a Hardware/Software solution for people that consume the products that creative make make. I have spent nearly 20 years on the Mac and now I am having to leave, if I want to do the things that I have been able to do in the past.

The ability to "eat your own dogfood" is going away. More importantly - 1 bad phone design and Apple becomes a much smaller company.

 

[ssgbryan]

If you don’t need or want MacOS, you can probably save money buying a PC. It’s been that way for decades.

PS Threadripper is PC class, and not sufficient or appropriate for the workstation market. EPYC is AMD’s equivalent to Intel’s Xeon. But you’re unlikely to see that on the Mac Pro, because Apple is unlikely to switch to AMD.

There are a number of companies making TR workstations.

How is it not appropriate? 64 cores/128 threads; faster I/O; greater IPC than Xeon; ECC memory; You only need to move to the EYPC AF series if you need more than 256Gb of ram.

 

[deconstruct60]

....

AMD’s 4000 series mobile chips offer TWICE the sustained performance of Intel’s offerings.

Source?

Twice (2x) is mostly arm flapping. It is more than a bit competitive.

+16% on Cinebench R20

The graphs on page 2 where relative percentage once put highly threaded workloads on. ( versus 9780H which pragmatically is a mainstream variant of the MBP 16" ) . Get up in 60% faster range. ( still short of that 100% faster though).

Ryzen 4000 performance benchmarks: Ryzen 9 4900HS beats Intel’s Core i9 mobile chips

AMD's Ryzen 4000 mobile CPUs finally break the curse that had the company losing to Intel in laptops over and over again. Our performance benchmarks prove it out.

www.pcworld.com

The fabrication process makes a significant contribution. Which is very relevant in putting some reality back into the "ARM is magically better than x86" claims or that "nobody can compete with Apple CPU designers" claims. Apple is catching Intel but they haven't necessarily got the tools to necessarily pass AMD. They aren't particularly going to out compete them on fab process for a substantive amount of time. ( Apple will get "months" earlier access because can wave more money at it (and pragmatically just buy up most of the wafer start slots. But that isn't technical prowess. )

Both AMD's and Intel's freedom of movement to upgrade to better fab is at least as good as Apple's. ( relatively Intel's is actually better since they have been still for so long. )

The 'baseline' that Apple's solution should have to pass at this point is largely AMD's , not Intel's, solutions. Desktop class, Apple's solution really aren't competitive at all. The "A series is desktop class" is often far more comparisons about the past, not the present ( or relatively short term future).

 

[Nütztjanix]

[…]
Desktop class, Apple's solution really aren't competitive at all. The "A series is desktop class" is often far more comparisons about the past, not the present ( or relatively short term future).

How would we know? Neither of us has seen this solution. It's all (more or less) educated guesswork, extrapolating from passively cooled, power consumption restricted mobile parts from phones and tablets.

 

[deconstruct60]

....
PS Threadripper is PC class, and not sufficient or appropriate for the workstation market.

That's nonsense.

A substantively high fraction of modern Threadripper motherboards come with ECC memory capabilities enabled and provisioned. The latest ones tend to come with better than 1GbE ports.

AMD classifies them as "desktop" processors but that is more so distinct from them being 'server room' / 'rack' environment targeted processors. Deskside workstations casually get dropped into 'Desktop' general category all the time.

Threadripper models don't sacrifice single threaded and max boosts speed as much as EPYC models do.
The both deliver in fundamentally the same chip package and use almost the same dies ( variation more so in Zen 2 on the i/O cihp than on the CPU chiplets. )

It has a different name than Ryzen for a reason.

EPYC is AMD’s equivalent to Intel’s Xeon.

It is equivalent to Xeon SP . Intel has several Xeon variants D, E , W , SP . EPYC is in no way shape or form competing with all of those.

Apple doesn't use SP, they use W. And AMD's competitive equivalent to

This vendor seems to think they are workstations https://www.boxx.com/systems/workstations/t-class
Certainly priced that way ( Xeon W and up class pricing. )

But you’re unlikely to see that on the Mac Pro, because Apple is unlikely to switch to AMD.

EPYC and Xeon SP are unlikely in a Mac Pro. Has little to do with a 'switch' .

 

[ssgbryan]

That's nonsense.

A substantively high fraction of modern Threadripper motherboards come with ECC memory capabilities enabled and provisioned. The latest ones tend to come with better than 1GbE ports.

AMD classifies them as "desktop" processors but that is more so distinct from them being 'server room' / 'rack' environment targeted processors. Deskside workstations casually get dropped into 'Desktop' general category all the time.

Threadripper models don't sacrifice single threaded and max boosts speed as much as EPYC models do.
The both deliver in fundamentally the same chip package and use almost the same dies ( variation more so in Zen 2 on the i/O cihp than on the CPU chiplets. )

It has a different name than Ryzen for a reason.

It is equivalent to Xeon SP . Intel has several Xeon variants D, E , W , SP . EPYC is in no way shape or form competing with all of those.

Apple doesn't use SP, they use W. And AMD's competitive equivalent to

This vendor seems to think they are workstations https://www.boxx.com/systems/workstations/t-class
Certainly priced that way ( Xeon W and up class pricing. )

EPYC and Xeon SP are unlikely in a Mac Pro. Has little to do with a 'switch' .

What is your definition of a workstation?

 

[firewood]

I‘m not saying it cannot be done, I’m asking what the timeframe is,

Likely similar to the timeframe that it took people to stop using 68k software, PPC software, and 32-bit software after those Apple architectural transitions started; if not shorter due to better development tools, and less assembly language coding these days.

That timeframe is typically shorter than most people's replacement cycle for their Macs.

So if you buy a new Mac now (and maybe a spare), then you likely won't need to buy another new different Mac until most of the software you need has been transitioned (and that new Mac runs significantly faster).

 

[ssgbryan]

Likely similar to the timeframe that it took people to stop using 68k software, PPC software, and 32-bit software after those Apple architectural transitions started; if not shorter due to better development tools, and less assembly language coding these days.

That timeframe is typically shorter than most people's replacement cycle for their Macs.

So if you buy a new Mac now (and maybe a spare), then you likely won't need to buy another new different Mac until most of the software you need has been transitioned (and that new Mac runs significantly faster).

So, buy 2 macs and hope that the software transition period is complete by the time the next Mac flops out the door?

That was over 2,000 days for the Mac Pro.

 

[PickUrPoison]

Every year it appears that Apple isn't the solution for more and more people that actually create. Apple has moved from a Hardware/Software solution for creative professionals/amateurs to a Hardware/Software solution for people that consume the products that creative make make. I have spent nearly 20 years on the Mac and now I am having to leave, if I want to do the things that I have been able to do in the past.

The ability to "eat your own dogfood" is going away. More importantly - 1 bad phone design and Apple becomes a much smaller company.

Ironic that just as Apple has the best range of Macs from low- to high-end there’s nothing for you.

iPhone is an excellent gateway drug to Mac (and iPad) though. In the earnings call, it was mentioned that around half of purchases of Macs and iPads were new to the platform. That’s no doubt why the installed base of Mac users continues to hit new records, seemingly every quarter.

Not sure who you think isn’t dogfooding. And yes, Apple releasing excellent iPhone after excellent iPhone is, undoubtedly, the reason that iPhone is half their revenue.
[automerge]1588477938[/automerge]

There are a number of companies making TR workstations.

How is it not appropriate? 64 cores/128 threads; faster I/O; greater IPC than Xeon; ECC memory; You only need to move to the EYPC AF series if you need more than 256Gb of ram.

Exactly. Apple’s not going to release a Mac Pro tower with half the memory capacity of an iMac Pro. I haven’t looked at the pricing lately but EPYC probably still isn’t cheap.

 

[PostApple2016BrainWash]

How would we know? Neither of us has seen this solution. It's all (more or less) educated guesswork, extrapolating from passively cooled, power consumption restricted mobile parts from phones and tablets.

If the new 12” or 15” ARM-Based Macs can be as fast as top of the line x86 CPU without increasing the TDP above 10w and passively cooled then it is really a good product improvement from Apple.

 

[09872738]

Ironic that just as Apple has the best range of Macs from low- to high-end there’s nothing for you.

I‘m sorry - what?
Apple really does not have a proper lineup at this point. There is a decent lineup of notebooks (lacking AMD CPU and nVidia GPU options though), but that‘s basically it.

- Mac Pro is grossly overpriced and put itself out of market for the vast majority of previous MP customers
- Mac mini upgrade paths ridiculously overpriced as well, in addition too locked down still (no upgradable SSDs)
- There is nothing in Apple‘s line-up in the desktop middle/upper middle class. No upgradable machine in the tradition of the cMP, the probably biggest market segment apart from mobile.

Apple at this point offers only Intel processors and AMD graphics, where AMD offers superior CPUs and nVidia still got the edge in GPUs.
I‘m sorry, but it sure looks like Apple does not take computers seriously any more, its computer lineup looks very half-hearted

 

[Nütztjanix]

If the new 12” or 15” ARM-Based Macs can be as fast as top of the line x86 CPU without increasing the TDP above 10w and passively cooled then it is really a good product improvement from Apple.

Even if they need to be 35W to be as fast or faster, it would still be a win for Apple. Because they'd be in control.

 

[PostApple2016BrainWash]

Even if they need to be 35W to be as fast or faster, it would still be a win for Apple. Because they'd be in control.

The purpose of using ARM-Based Macs is to provide better performance and use less power consumption than x86 like 10w instead of over 30w on a portable device.

 

[Nütztjanix]

The purpose of using ARM-Based Macs is to provide better performance and use less power consumption than x86 like 10w instead of over 30w on a portable device.

Certainly one point to consider, but surely not the only one.
And, depending on which Mac you look at, 35W would be a significant improvement in power consumption (16" MacBook Pro, Mac Mini for example). Others not so (13" MacBook Pro, MacBook Air).

 

[ssgbryan]

The purpose of using ARM-Based Macs is to provide better performance and use less power consumption than x86 like 10w instead of over 30w on a portable device.

No it isn't - It is about rebuilding the walled garden, since Apple no longer competes in the PC marketplace.

 

[futureisfilm]

I‘m sorry - what?
Apple really does not have a proper lineup at this point. There is a decent lineup of notebooks (lacking AMD CPU and nVidia GPU options though), but that‘s basically it.

- Mac Pro is grossly overpriced and put itself out of market for the vast majority of previous MP customers
- Mac mini upgrade paths ridiculously overpriced as well, in addition too locked down still (no upgradable SSDs)
- There is nothing in Apple‘s line-up in the desktop middle/upper middle class. No upgradable machine in the tradition of the cMP, the probably biggest market segment apart from mobile.

Apple at this point offers only Intel processors and AMD graphics, where AMD offers superior CPUs and nVidia still got the edge in GPUs.
I‘m sorry, but it sure looks like Apple does not take computers seriously any more, its computer lineup looks very half-hearted

Apple doesn’t take their computers very seriously?

Apple still has a low in computer. The Mac Mini, while not $499 is still a good bargain. Linus Tech Tips review of it was pretty favorable even if there was some caveat’s.

Then the iMac line up adds value to the Mac line up. Having one of the best displays and still one or the best All in One’s available.

The iMac Pro is the all around pro machine for those pro’s looking for something quick and powerful enough for 4K video editing or other pro applications.

The Mac Pro is meant for huge corporations who are looking for something more customizable and powerful enough for their heavy workloads.

And their laptops are getting replaced with better keyboards and the 16 MBP has better thermals.

And not to mention Apples iPad and Mac’s made a lot of money last year. They aren’t releasing their numbers but they are releasing how much money they made and it’s more then what Dell and HP made.

So yes, they are serious about their computers. But they don’t cater to every use case and for whatever reason they still have an exclusive deal with Intel processors. If they weren’t able to make their own processors then they might switch to AMD. But, for a Whole myriad of reasons, they seem focused on getting that ready. We don’t know what that looks like just yet. And until we do, we can’t speculate on exactly what their plans are for the Mac’s.

 

[ssgbryan]

Ironic that just as Apple has the best range of Macs from low- to high-end there’s nothing for you.

iPhone is an excellent gateway drug to Mac (and iPad) though. In the earnings call, it was mentioned that around half of purchases of Macs and iPads were new to the platform. That’s no doubt why the installed base of Mac users continues to hit new records, seemingly every quarter.

Not sure who you think isn’t dogfooding. And yes, Apple releasing excellent iPhone after excellent iPhone is, undoubtedly, the reason that iPhone is half their revenue.
[automerge]1588477938[/automerge]

Exactly. Apple’s not going to release a Mac Pro tower with half the memory capacity of an iMac Pro. I haven’t looked at the pricing lately but EPYC probably still isn’t cheap.

Best range my !@@.

Eypc is about 1/3 the price of Intel's equivalent - well not exactly - Intel doesn't have anything equivalent to the top end Eypc CPUs.

Mac Mini - Tops out at 6 cores, 32 Gb of ram (which wasn't enough in 2007) - integrated Intel graphics for over $,1000. No ability to upgrade. Added bonus - thermally throttled.

iMac - $1,100 gets you a 2 core/4 thread system (in 2020!) 8Gb of ram, a 1Tb of spinning rust, integrated graphics. Maxes out with a 6 core mobile chip and a cut down Polaris 560 (No longer manufactured - two generations back) GPU. 64Gb (which wasn't enough in 2009) of ram is a $1,000 option on purchase. Can add ram, if you are willing to disassemble the entire computer. Added bonus - Screen Roulette. Added bonus - Thermally throttled.

The 27" iMac maxes out with 6 cores (less than I had in 2007), 64Gb of ram (less than I had in 2009), a cut down Polaris card from 2016, spinning rust for hard drives (2Tb max). The screen is nice.

iMac "Pro" - Maxes out with 18 cores/36 threads @2.3Ghz (before AVX-512 offset, if your software uses that instruction set), a Vega 56 gpu (no longer manufactured, since it is last generation), all of which is thermally throttled. It is good that it can reach 256Gb of ram, and has HBM video memory, but all of that will run you over $14,000.

Or about 3 times the cost of an equivalent Threadripper box.

Mac Pro - $6,000 for an 8 core machine. Crushed on the low end by Ryzen 3950x systems, Crushed in the middle by Threadripper, Crushed at the top by EYPC AF series of processors. All for 1/4 to 1/3 of the price.


Futureisfilm - nothing in the Apple lineup is a "good bargain". Everything outside of the Mac Pro is thermally throttled - you literally can't push either the CPU or the GPU and stay at stock clocks. Maxing both is simply out of the question. They are massively overpriced and are full of obsolete hardware to boot. $1,100 for a 2 core system.

The Mac Pro is $1,200 dollars worth of parts in a $4,700 enclosure.

 

[deconstruct60]

Even if they need to be 35W to be as fast or faster, it would still be a win for Apple. Because they'd be in control.

"Control". The T2 chip already doesn't allow the main x86 CPU to see the master version of the firmware ( it only gets a copy). How much more control does Apple need? They have parameterized and control the boot process.

It isn't about control for control sake.

Apple has a desire for thinnest, lighest possible all metal laptop ( phone , etc.) that spans beyond what Intel/AMD are set up to do. (and most of the PC market for that matter). Pursuing that 'holy grail' they have an inside track. If Apple thinks an Apple CPU can add major, differential value to one of the Mac products they make then can see them perusing that with some narrow set of resources.

But the entire Mac line up is about 8-12 different CPU configurations. There are just about more CPU products there then in the rest of Apple whole product line up. [ majority if A-series powered products are using "hand me down" packages. That extremely points to not wanting to do more CPU products for individual products. ] .

Lots of work? Apple is usually pretty keen on farming that out. Making products? subcontracted. Displays ? subcontracted. RAM ? Subcontracted. etc. etc. etc. Apple doing everything for everybody runs completely antithetical to modus operandi over the last 20+ years.

One T-series that got inserted into the whole non ARM Mac line up means they could probably get by with 1 ,2 ... maybe three ARM chips for the whole line up. 1 ( maybe 2 ) low end laptop chips ( could also do a sub $600 Mini replacement) and couple the rest to a T-series follow on.

 

[chucker23n1]

"Control". The T2 chip already doesn't allow the main x86 CPU to see the master version of the firmware ( it only gets a copy). How much more control does Apple need? They have parameterized and control the boot process.

Control over SKUs and their schedule.

 

[Nütztjanix]

"Control". The T2 chip already doesn't allow the main x86 CPU to see the master version of the firmware ( it only gets a copy). How much more control does Apple need? They have parameterized and control the boot process.

As @chucker23n1 said - it‘s not about control over what you can or cannot do. It’s about control when and what chips they stick into their machines. I guess they’re very much tired of waiting for Intel to get their game together.
Switching to AMD might seem like an option right now, but as @cmaier already pointed out, they don’t have a track record of keeping an advantage over a longer time span.

 

[deconstruct60]

I am going to have to break this post down. Because it's full of odd assumptions. I understand this entire thread is all about assumptions and opinion on Apple switching their Mac's to ARM, so having assumptions is to be expected.

However ... lets begin. Why do you think the Mac Pro had little effort into it? The Mac Pro's Intel Xeon was released in 2019. Sure its not more powerful then the AMD EPYC and Ryzen processors that came out last year, but it was some of the best Xeon processors last year -- which the Mac Pro came out last year, so that's saying something.

Apple put little development effort into the Mac Pro from 2013 to 2017 because strategically it didn't make much of a difference. Apple grew the Mac market and ecosystem all through that period. It is a "nice to have" product, not a strategic one. Same kind of hand waving that it was strategic is similar to the assertions bandied about when Apple kill the XServe ( Apple will die without a real "pro" server room options. Mac Server will go down etc. And yet didn't implode at all and Mac Server sales went up for a substantive amount of time. )

If Apple had a team that actually worked on the Mac Pro deliberately they could have done one in 2017-2018 along with the iMac Pro . They didn't.

I'm not talking about the last end phase of Apple's Rip van Winkle span. From mid 2017, most of 2018 , and early 2019 they probably did put substantive effort in. But in no way shape or form did it require 3-4 years to get to the point they did with the Mac Pro. That should have been about 18 (or less ) months of work for a reasonably resourced development team. In 2017, they were doing "dog ate my homework" sessions and didn't have anything substantive to show at all. That was low effort.


The Mac Pro 2019 is work, but it isn't timely.

And we know they at the very least started development of the Mac Pro in 2017. Probably before.

Probably not much before. In the 2017 "dog ate my homework" session they mention that looking to do something in the iMac "Pro" space. And iMac Pro showed up later. Product manager for Mac Pro and iMac Pro the same ( PM bio description here. )

Evident also in the rest of the Mac line up also. If the Mini has update than iMac probalby isn't doing much. Move one laptop forward ... another probably gets a minimal update. Apple hasn't done a "walk and chew gum at the same time" , across the board upgrade to the Mac product line up in more than 6-7 years. "Well that is all Intel's fault" is rather lame excuse. it isn't all Intel's fault. Apple has poured major reasources into othe areas where has "cost managed' the Mac Product space relatively closely.

So that was TWO years of development. How is that not putting a lot of time and effort into the Pro machine they made?

So if Apple hits the snooze bar and goes back to sleep for another 2-3 years before starting development again that will be putting in effort? Not really.

With those two years on Mac Pro the iMac Pro progressed how? Pretty likely they are now doing something with it, but that also probably means they are not on the next Mac Pro.

Apple getting a the W5700X out within 6 months of getting the Mac Pro out is a partial sign that perhaps they are not fully going back to sleep.

And this is the primary point. if Apple can't bring itself to put a full time team on the Mac Pro ( or Mini ) why are they going to put a full time team on a processor for said products. Letting Intel (or AMD) do all the CPU work is actually easier. That is the huge disconnect in what is being pitched here about multiple CPU packages for all Macs.... Apple is going to throw tons of custom silicon work at products that are order ( if not orders in Mac Pro case) magnitude smaller volume than the iPhone .

I don't think Apple plans on making one desktop for all of their products. I think each line will have their own processor, but slightly tweeked. Because Apple likes to design for each product and based on the needs of that product.

They didn't even put effort into the overall products like that when didn't have to do the processors work to do for them. Why are they doing a 180 move on that now? Most often that question is answered by pointing at the Scooge McDuck money pit with billions in cash in it. That isn't a creditable answer. That is how Apple can spend money; not make money.

Yeah, Apple could have made a generic tower and throw some random parts in and not tried to make a high quality product. But they didn't. They thoughtfully designed a case that is easy to upgrade and add parts into the tower.

Right! Which is even easier. And yet all through 2014-2016 did a whole lot of nothing on the Mac Pro. So why is some multimillion dollar constant effort going to be put a Mac Pro processor ?????????? They thoughtfully designed case costs probably isn't near what the chip design costs will be.

And there is a pretty goods reason to design their own Pro Processor. They are switching to ARM, and unless Apple is completely changing the wheel and supporting Intel and Arm Mac's forever they can either never make MacBook Pro's, iMac Pro's and Mac Pro's ever again OR they can make an ARM processor.

Apple doesn't have to . They could wait for someone else to do it. Apple isn't the sole possible answer in the ARM space any more than Intel is the sole possible answer in the x86 space.

Apple also doesn't "have to " do a 'big bang' transition to ARM either. If Apple used ARM to expand the the Mac ecosystem range they could get back into the sub $700 space in laptops and desktop and grow Macs out of 6-7% share range into something closer to 10%.

x86 prices are coming down. The biggest "big picture" thing about the Comet Lake is not the 5.3GHz or ten cores... it is the price competition building between Intel and AMD.

Intel Comet Lake-S Arrives: More Cores, Higher Boosts and Power Draw, but Better Pricing

Kick the tires and light the TDP fires

www.tomshardware.com

The Mac Pro (and also individually the iMac Pro) are probably pretty close to being in the sub 100K/year run rate zone. Where is the business case for Apple only ARM processors for those sorts of volumes? General x86 Workstations market numbers in the millions per year. Apple is an order of magnitude down from that in volume.


The entry CPUs in the Mac Pro (and iMac Pro) doesn't cost "Thousands". The folks who arm flap that Macs are going to get soooooo much cheaper then Apple goes to ARM . There isn't much to that for most of the Mac product line up because the volume isn't there at all.

They might take an actual design like the Neoverse N1 and customize it to their liking or they can do what they have been and design one from the ground up. Which is kind of perfect, because the Mac Pro won't be due for an update until 2022 or 2023 so they have several years to test and develop Mac Processors leading up to their Pro Level processors.

Maybe N2 (or N3). That is possible. But if someone else is doing that and Apple can split the development costs with some other system vendors that would be even lower cost for Apple. The design isn't the issue. If the volume of the Mac products as you go up the product line to the higher priced products. There isn't much there to support chip design , validation , variation that is highly detached from what is going on in the "customized for phones" design. ( doubtful going to have 128GB RAM phones any time soon or even intermediate term. )

One major reason Apple consistently stays ahead of Qualcomm and some of the others is that Apple typically does on CPU design per year. ( even the X series isn't on the 12 month cycle. They are on process shrink cycles which a 18-24+ month cycles. ). Qualcomm does about 6-8 processor variants a year. At least a few of those are a necessary distraction because they are trying to be everything for everybody in the phone space. That is a dual edge sword tactic. The counter edge is that probably can't catch Apple as long as do that. ( simulator time, resources , etc for those other products that Apple could take equal allocation to improving their single product. )

So I am not sure why you believe they have no business reason to continue to make Pro Machines but Apple will continue to make them. I seriously doubt they will just forget this segment and never make another Pro Machine again.

A business reason is where they make a profit at it. Few of the hand wavers in this thread have pointed to how the overall development costs and resource needs go down as you produce more variants of a CPU microarchietcture. That is because the don't. There is volume threashold have to hit to make doing a custom processor that makes business sense. That is highly lacking in these "ARM for everything" threads.

If Windows 10 picked up ARM desktops that probably would be at least as big as what the Mac market is. Together they make something viable. But Apple's sub 10% share and then looking at the upper end Mac share of that 1-2% of 6-7% that isn't a business case.

Similar short sighted thinking goes into Apple taking on the modem development. If iPhone sales flatline Apple over extended period of time won't be saving much over what were paying Qualcomm. There is a dirt load of work that pulling in to track all the future standards and interoperability development overhead. ( won't be surprising to see Apple push modems into other products to get volume up at that point. .... but that synergy doesn't bode to well with their desktop products. )



P.S. How does Amazon make money on Graviton 2 when they are probably in the highly sub millions run rate? They aren't selling them. They are renting them. Lower power means lower operating costs which can in part be allocated to paying for the chips. Let alone the "rent" for the computation cycles. They are being thrown at very high margin x86 sever CPU alternatives. Easy to look thin when standing next to the circus side show fat lady. If the AMD and Intel price war continues it won't necessarily turn out so well for future Graviton iterations.