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

[laz232]

ARM processors are significantly cheaper than Intel's. Obviously, Apple would have to offer a lot more cores than Intel and this will have a cost but I don't expect to see any difference up or down by moving to ARM.
There was no difference when they moved from 68000 to PowerPC or when they moved from PowerPC to X86, and then from X86 to X64. The A13 is apparently worth 15 bucks. That's the same as Qualcomm 865. Processor costs won't influence the value of the Hardware. They're such a small part of BOM.

They will increase the profit margin for Apple. The Intel chips are - rightly or wrongly - much more expensive. Apple likes a high profit margin - if there is a typical 3xBOM mark-up then they have an extra 200+ USD of profit margin and can amortise A series dev costs over the full range of products.

Like I said though - it will mess up my use of the machines as I donæt use iOS apps - but need full speed Windows in parallel to macOS
[automerge]1587666694[/automerge]

I wouldn't be surprised if Apple did go to ARM that they also use an intel x86 CPU onboard for, wait for it... X86 required executions! This is the most likely scenario IMO and could be a BTO option. Think of it, ARM for everything MacOS and everything that ARM can run natively and then the secondary x86 CPU for Bootcamp, etc, etc, etc... The cost of Intel (or AMD for that matter) CPU's in bulk is super cheap for Apple. You also wouldn't need the biggest baddest x86 CPU either.

You're completely wrong on many levels. Is not going to happen.

 

[ph001bi]

I can only speak for my company, where we have both macs and PCs. People get macs to run macOS, and not windows - at least in my organization. Its actually a poor use of money for a company to purchase a Mac just to run windows.

In my company where Macs are used for development there are a few people who need Windows for one or two apps. Some use Parallels others use Citrix VDI (I am one of them). We have 12000 devices, 500 of them are Macs, and maybe a tenth of those at a stretch need some kind of Windows solution. We're talking about a niche of a niche. If Apple sells 4.5M Macs instead of 5 millions in a quarter because of the switch, they won't even notice it.

 

[smulji]

Yeah. Was it a full on transition last time? They just gradually phased out PPC hardware.

Azrael.

It wasn't gradual. Apple phased out PPC in 1.5 years.
[automerge]1587667370[/automerge]

I'm super late to this thread - but this is what I was thinking. What better way to show confidence in your decision than to release the next iMac redesign with the ARM chip.

According to rumors, the redesigned 23" iMac is coming out this year and ARM Macs are coming out 2021 so that doesn't exactly align

 

[ssgbryan]

The real question is was it worth it in the end? Would it have been better if we stayed on PowerPC?

Hardware wise, yes. I would have liked to have had a 5Ghz (8 cores on a dual CPU board) Power Mac in 2007. Today, we could have had 4 threads per core (or 8 threads per core on the expensive chips) on a Power9 based Power Mac.

7nm Power10 systems inbound next year.
[automerge]1587667545[/automerge]

And everyone will say it's a catastrophe because _____ (fill in the blank with an app from Adobe) hasn't been ported.

If it happens.

A lot of software didn't make the transition from PPC to Intel.

Get ready to rework your workflow, and you won't be getting that software for free.

 

[Zdigital2015]

Has anyone seen Thunderbolt work in an ARM system before?

Thunderbolt 3 support is simply having a support chip (e.g. Intel JHL-7540) that connects to the PCIe lanes hanging off of the CPU or the PCH. Apple's T2 handles many of the functions of a PCH and Apple's ARM/Arm CPUs will most likely support PCIe 3 or 4 natively on the CPU die itself. If they don't, I'm sure they've come up with something else. Remember they used to build their own computers using PowerPC chips that had PCI and PCI-X slots, I'm not sure why anyone thinks that the next go round is going to be different or that Apple doesn't know what they're doing.

 

[Orionfox]

Yeah, real gamers game on PC. It is also more cost-effective and you will have a better experience as well.

I guess I am not a "real" gamer then. BTW, anyone who uses the term "real gamer" isn't my kind of gamer. I am a "Computer gamer". I play RPGs only anymore. I only play with keyboard and mouse. I ONLY play on a computer - my Mac.

What do I know, I haven't stopped gaming since the early 80's with my Vic 20 and Atari 2600. I did drop consoles in the very early 90s except to play Nintendo with my kids.

 

[dannyyankou]

Hardware wise, yes. I would have liked to have had a 5Ghz (8 cores on a dual CPU board) Power Mac in 2007. Today, we could have had 4 threads per core (or 8 threads per core on the expensive chips) on a Power9 based Power Mac.

7nm Power10 systems inbound next year.

Steve Jobs said they couldn't even achieve a 3ghz PowerPC, I don't think 5ghz was happening.

 

[cmaier]

When Apple switched from 68k to PPC, and from PPC to Intel, they were fighting for their existence. Currently Apple is the highest market cap company on the stock market, it’s a different situation. They have nearly endless resources, they don’t need to put all their eggs in one basket.

They might not have to, but they will. There are zero instances of them ever doing the opposite. Heck, they don’t even support AMD cpus or NVIDIA GPUs. They *could* do those things, but they don’t.

 

[AHDuke99]

I really worry we will see a return to the 1990s or early 2000s when a lot of software was not available on the Mac or we had to wait 6-12 months for a port of the windows version. I'm sure iOS apps will easily run, but for power apps and games, it's going to be a painful move.

 

[ssgbryan]

I guess I am not a "real" gamer then. BTW, anyone who uses the term "real gamer" isn't my kind of gamer. I am a "Computer gamer". I play RPGs only anymore. I only play with keyboard and mouse. I ONLY play on a computer - my Mac.

What do I know, I haven't stopped gaming since the early 80's with my Vic 20 and Atari 2600. I did drop consoles in the very early 90s except to play Nintendo with my kids.

You have missed out on a lot of fun games, then.

 

[davepete]

Many commenters here think Apple would replace an Intel chip with a slower or equal-speed A-chip. My prediction is Apple will replace Intel chips with 2x faster A-chips. It's a good marketing claim: "Priced $200 less, now twice as fast." Once their low end A-chip Mac laptops have better performance than their still-Intel Pro laptops, users will be begging them to switch the Pro models too.

As for compatibility, I actively write and publish more apps than you would believe on the App Store, a few on the Mac App Store, using Apple's APIs, no 3rd party frameworks other than Google AdMob. I have a feeling some macOS developers using AppKit will be left behind, just as the Carbon macOS developers were. But, frankly, most new software I use is being written for iOS, and it's a cinch to port iPad apps to macOS now (you click a checkbox). The Google AdMob frameworks are a no go, but Apple's latest APIs are all fine.

If Apple can expand the macOS user base by lowering prices and doubling the speed, devs will click the Mac checkbox and publish for it. For all I know, Apple may let existing iOS apps run natively on an A-chip Mac, clicking the checkbox or no. And when we submit Mac-only apps for review on appconnect now, I imagine Apple is already recompiling and running them on A-chip Macs. I think the ability to run millions of iOS apps natively on macOS would far outweigh running Windows emulators. (Bootcamp and Parallels are tedious and finicky.)

 

[sbarton]

do you really believe that Apple hasn’t thought about this issue?

Remember the PowerPC to intel transition. It was a pain in the ass yes for 1-2 years, but after that it’s a much better strategy for them.

Well, actually it was 10 years of photoshop "bake offs" to try and convince us that PPC was faster than Intel as well as all of the cost and aggravation of a dreadful ecosystem of a non-intel architecture coupled with proprietary graphics hardware and walled-garden APIs.

...and THEN 2 painful years of transition.

Now that SJ is gone it's like this company feels the need to RUN, not walk back to the past.

 

[theluggage]

Yeah. Was it a full on transition last time? They just gradually phased out PPC hardware.

Well, hardware-wise, PPC to Intel was done and dusted in slightly over a year, so, yeah, it was full-on... but in that case, PPC development had completely hit the buffers with no G5-class chip available for laptops so it was kinda urgent, and the new Intel Core processors were a big speed bump... (I think the initial skepticism was that Core hadn't been announced and the thought of a Pentium 4/Netburst space heater in a Mac made people throw up a little bit in their mouths).

I think an ARM transition would need to be a bit more sedate - there won't be such a big step-up in speed to make up for the initial need for emulation. Plus, if Apple totally discontinues the new Mac Pro in favour of an ARM version anytime this side of 2024 I think the remaining Pro users are going to leave the building...

 

[ssgbryan]

I really worry we will see a return to the 1990s or early 2000s when a lot of software was not available on the Mac or we had to wait 6-12 months for a port of the windows version. I'm sure iOS apps will easily run, but for power apps and games, it's going to be a painful move.

Moving to windows wasn't hard at all.

AFA Hardware, Everything just works....

AFA Software, So much more......

 

[NickName99]

Well, hardware-wise, PPC to Intel was done and dusted in slightly over a year, so, yeah, it was full-on... but in that case, PPC development had completely hit the buffers with no G5-class chip available for laptops so it was kinda urgent, and the new Intel Core processors were a big speed bump... (I think the initial skepticism was that Core hadn't been announced and the thought of a Pentium 4/Netburst space heater in a Mac made people throw up a little bit in their mouths).

I think an ARM transition would need to be a bit more sedate - there won't be such a big step-up in speed to make up for the initial need for emulation. Plus, if Apple totally discontinues the new Mac Pro in favour of an ARM version anytime this side of 2024 I think the remaining Pro users are going to leave the building...

Apple doesn’t have to go all-in with only one or the other, that’s a risk they don’t need to take. They’re under no pressure, why bet the farm?

MacBook Air should use an Apple processor, way better battery life and way better performance. Pro models? Let the customer choose.

Who knows where Intel and Apple processor performance will be years down the road. Apple can just use whichever processor makes the most sense in each product, and/or let the customer pick if that makes sense.

 

[Hyperchaotic]

It is different, when they moved to X86 they didn't need developers to get onboard, Apple were getting onboard into huge world that already existed.

I remember that transition well, you are wrong. They had to include PowerPC emulation "Rosetta". They spent a lot of effort getting developers to port to x86 and many developers said they'd abandon the platform with the switch to x86, some did and others came back later. Getting developers onboard were definitely a big issue with the PPC->x86 transition.

I'm guessing they're thinking it'll be easier this time, with so much stuff being web/cloud based and even more high level languages being used. Chromebooks seem to be doing well, so a transition starting with similar type lightweight laptops would be one way to get going.

 

[Donar]

Mac mini ARM please. I already have a G4 1,5 GHz PPC, Intel 2009 Core2Duo, 2012 i5 and 2018 i5 mini, so the more the merrier.

 

[theluggage]

Many commenters here think Apple would replace an Intel chip with a slower or equal-speed A-chip. My prediction is Apple will replace Intel chips with 2x faster A-chips.

Bear in mind that the iPad Pro A12X benchmarks put it on a par with the 2018 i7 15" MacBook Pro.

Now, I'm healthily sceptical about the chances of it beating an actual i7 15" MBP in real life, but I'd say that was more than good enough for a MacBook Air/12" MacBook type of machine, which is going to be where the ARM has the initial advantage.

Developing an ARM-based Xeon-W killer might take a little longer but I wouldn't bet against it in a few years (there are 80 core ARM server chips although they're not optimised for workstations - I'm thinking something with on-chip Afterburner-type tech would fit the bill for a future Mac/iMac Pro) .

 

[nol2001]

I wonder how they are going to market the MacBook-esque model.

Apart from better battery life, weight and possibly price, what else?

I’m hoping the screen is 120fps, there will be no text under the display and definitely no touch bar.

 

[NickName99]

They might not have to, but they will. There are zero instances of them ever doing the opposite. Heck, they don’t even support AMD cpus or NVIDIA GPUs. They *could* do those things, but they don’t.

Do you really think this is anything like offering either Intel or AMD processors, or offering either NVidia or AMD GPUs? Those are just competing implementations, jockeying for the lead year after year. If it made financial sense for Apple to offer a choice there, they would have.

This is an unprecedented situation. Apple can do this any way they want, and rest assured nobody in this comment thread has come up with anything Apple hasn’t considered.

 

[Zdigital2015]

I really worry we will see a return to the 1990s or early 2000s when a lot of software was not available on the Mac or we had to wait 6-12 months for a port of the windows version. I'm sure iOS apps will easily run, but for power apps and games, it's going to be a painful move.

This is not the 1990s...at that time, there were serious questions for developers who wondered if the platform was dead. Now, the platform is doing well and is poised to d even better once the ARM/Arm transition begins. Apple has to maintain and expand the synergy between devices (iPhone, iPad and Mac) to make it an even more seamless transition to work between your preferred platform and the other devices you own. That’s the real trick, not moving to ARM/Arm.

 

[Bruno Castelló]

Long-term, though, there's no reason why everything shouldn't go ARM.

Here's my reason, based on my experience with ARM based computers (not smartphones):

They shouldn't, because ARM is s**t. End of story.

If all the phones (not only iPhones) have battery drain issues with ARM chips, then a portable computer like the MacBook Air with a 12" screen will have BIGGER issues. Mark my words. Apple always lies about their battery durability when they launch a new phone or macbook.

 

[cmaier]

Do you really think this is anything like offering either Intel or AMD processors, or offering either NVidia or AMD GPUs? Those are just competing implementations, jockeying for the lead year after year. If it made financial sense for Apple to offer a choice there, they would have.

This is an unprecedented situation. Apple can do this any way they want, and rest assured nobody in this comment thread has come up with anything Apple hasn’t considered.

No, i don’t think it’s ”just like.” I think it’s WAY EASIER. And yet they don’t do it.

And it’s not an unprecedented situation. They went from x86 to x86-64 and had plenty of money in the bank to keep supporting everyone who was crying about not being able to continue running 32-bit applications. But they ousted 32-bit.

They could have kept supporting usb-a. They didn’t.

When the decide to change technologies, they do it. Every. Single. Time.

There are zero counter-examples.

And supporting two completely different architectures is way harder than these other things. Within a couple of years, there is zero doubt that it will be all ARM, other than maybe the Mac Pro which, due to its very slow update schedule, may dangle by a thread for awhile longer.

 

[dugbug]

But since A series are portable chips, presumably it will have a battery? Having to plug an A series Mac into line voltage seems very fetishy. It’ll have USB C for power, and a battery. My long time dream of a portable battery powered Mac mini, with wireless display and keyboard could be around the corner! Say goodbye to the compromises of trying to put a hot processor in a laptop!

well appletv and homepod are A devices that are wall-powered with no battery but the idea would be power efficiency gains and maybe fanless design

 

[cmaier]

Here's my reason, based on my experience with ARM based computers (not smartphones):

They shouldn't, because ARM is s**t. End of story.

If all the phones (not only iPhones) have battery drain issues with ARM chips, then a portable computer like the MacBook Air with a 12" screen will have BIGGER issues. Mark my words. Apple always lies about their battery durability when they launch a new phone or macbook.

What Arm based computers?

I’m a cpu designer. Arm is fine. It’s no more or less capable than SPARC, x86, MIPS, PowerPC, or any other instruction set architecture.

Please show me where apple has lied about battery durability (or, for that matter, made ANY statement about battery durability) when launching a new device. You can’t.