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

[Spectrum]

I have a (currently irreplaceable) piece of software that was last written in 2001 for PowerPC. But I can still run this in 2020 on Mojave/Catalina via a VM running OSX 10.6.8.

Hopefully, any ARM transition will retain the ability to virtualise prior OSX operating systems...

 

[neuropsychguy]

Yep.. By that logic the Mac should've been dead a long time ago.

You do realize that Apple has completely switch processor architectures serveral times already in the past right?

This isn't Apple's first rodeo.

It is the first rodeo for some (many?) who comment here. Things can look like catastrophes without an understanding of history.

 

[timber]

I am quite excited about these new ARM Macs.

A14s or A15s and beyond without the thermal (and ohter) constrains of phones or tablets are exciting. And the large number of iOS devices sold give Apple all the incentives, particularly financial, to keep them on the cutting edge. And while it is true that x86 is the standard "serious OS" platform, ARM is the standard "everything" platform so simply it isn't fair to say that Apple are setting themselves for another "Power PPC" lone walk moment.

Those SoCs also have pretty good GPUs in them particularly when compared with what intel puts out. Every Mac will graphically quite good while nowadays most of them are particularly poor. dGPUs are still available for the rest of the lineup. I do expect Apple to just stop relying on both AMD or NVIDIA.

Most people won't even notice just like most people don't know where the engine of their car is and which sets of wheel is it connected to.

 

[azentropy]

millions of developers code for ARM, using x86 machines. You can test on a different machine than you develop on. It will be fine.

Lol, not in our environment/company.

 

[Azrael9]

The Mac market is about 20M per year. 10% is 2M systems per year. Over 4-5 years that 8-10M users.
It is indicative to a highly contributing reason why there is a decent chance won't flip the whole line up to ARM.

2M $2K systems is leaving $4B on the table. In the grand picture of Apple overall revenue ... maybe "no big loss". In the Mac product space, it is.

Apple phased out Mac OS9 and PPC to the new Intel platform. They'll phase out what we have now to the new ARM platform is it suits them.

Azrael.

 

[Bruno Castelló]

You are right. But I am not worried. Late 2020's would probably so different that all our prior assumptions right now are moot.

But I will stick with x86 for the time being. Because I know it will take time to port apps to ARM, if ever.

I am not willing to compromise my productivity to be a "beta tester" for the first few years.

Same here. But I won't be helping to port anything to ARM. I have zero faith in ARM based computers.

 

[Jyby]

Science Rules!

 

[Azrael9]

if Apple hasn't given them a sniff, then why AMD references where found in last macOS code? Seems they have been testing AMD for their OS.

That's a good point...

Azrael.

 

[jlc1978]

This is no use to anyone relying on x86 programs, though. The Windows on ARM version runs x86 programs in emulation and it's slow as molasses.

It doesn't matter if the OS runs on ARM; it matters if the programs required are x86-only.

Exactly. Just because it is Windows doesn't mean everything runs on it acceptably. I use Parallels to run PowerBI as it has no Mac equivalent. I really don't want to have a separate WinBoxen if the Mac goes all ARM.

However, to me the question is, once OS X transitions to ARM will it run on the iPad?

 

[Bug-Creator]

The same iPad Pro that cant run an emulation of Windows 10 with desktop speed? Yeah, right.

*shrug*

Sofar noone really made ARM chips >20W as there was no demand, but the reality is that with the right amount if R&D the same process used and the same target TDP the instruction matters very little in the resulting (raw) performance.
So x64, ARM or even an 68000-64 would all run at similar speeds.

Performance lost be emulation can be anywhere between 30 and 95% depending on what actually is emulated (just a CPU or the whole system) and how it is implemented.

No reason why Apple couldn't come up with a >200W ARM-CPU for the 2025 MacPro and have it be really competitive.

 

[Azrael9]

I am quite excited about these new ARM Macs.

A14s or A15s and beyond without the thermal (and ohter) constrains of phones or tablets are exciting. And the large number of iOS devices sold give Apple all the incentives, particularly financial, to keep them on the cutting edge. And while it is true that x86 is the standard "serious OS" platform, ARM is the standard "everything" platform so simply it isn't fair to say that Apple are setting themselves for another "Power PPC" lone walk moment.

Those SoCs also have pretty good GPUs in them particularly when compared with what intel puts out. Every Mac will graphically quite good while nowadays most of them are particularly poor. dGPUs are still available for the rest of the lineup. I do expect Apple to just stop relying on both AMD or NVIDIA.

Most people won't even notice just like most people don't know where the engine of their car is and which sets of wheel is it connected to.

Some sound points. Especially re: GPU.

Yes. Life will go on.

Azrael.

 

[Jyby]

Science Rules!

 

[ani4ani]

The ever-reoccurring “Apple is doomed” is part of the Macrumors spirit. It will never go away.

In the old parlence, they only have to be right once, those who don't believe it have to be right everytime

 

[ericwn]

Apple finally is tired of 14nm+++++++++++++++

I would get that keyboard fixed!

 

[123]

if we could only count all the times “Apple is doomed” has been predicted because of a change in processor architecture… 😅

Well, they were doomed. Luckily they switched to Intel just in time.

 

[now i see it]

Don't fret chaps, undoubtedly there'll be a Rosetta 2.0 workaround so that all the "legacy" x86 code will run fine with a small performance hit

 

[Jyby]

Science Rules!

 

[Bruno Castelló]

Don't fret chaps, undoubtedly there'll be a Rosetta 2.0 workaround so that all the "legacy" x86 code will run fine with a small performance hit

GOD FORBID THAT!

 

[Azrael9]

I'm curious what percentage of current Mac owners actually boot into Windows. It certainly is a very convenient feature, but if only 10% use it, it's not a big loss. I'm guessing Apple must have some metrics on this before 'walking away' from this.

I used BootCamp for a very old game. (One my iMac gpu got fried playing, it seems...)

I might as well just buy a PC for that.

Mac going ARM won't affect me. If a redesigned iMac comes this year. I'll buy it. And be able to run my apps for years to come.

Plus, buying that PC. You can run a Hack'tosh on it with careful selection of parts.

Any ARM Mac on laptops won't phase me. I don't like laptops. It'll happen. It sounds very exciting. The rise of ARM on iOS has been intriguing to watch.

We'll look back on this thread as quaintly as the PPC transition. There will be a fuss, a bit of ruckus and then it will be eventual resignation or buy a Windows PC or some such.

Azrael.
[automerge]1587651776[/automerge]

GOD FORBID THAT!

Yes. But God (jobs) didn't forbid that.

Azrael.

 

[123]

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.

You know why it was much better? Because they aligned with the main stream with all the benefits that came with it once the transition was complete.

This time, it will be a pain in the ass for 1-2 years for the transition. And then it will continue to be a pain in the ass because nothing will be compatible.

 

[Azrael9]

Well, they were doomed. Luckily they switched to Intel just in time.

Yes. It did more for Marketshare than PPC did.

And iOS drawfs the Intel Macs in circulation.

It's not like we can't see the 1 billion and counting train coming down the track.

iOS expansion into other 'Mac' like devices is inevitable. We have watches, pads, phones...

Azrael.

 

[bernoulli]

Newsflash: most software most people rely on is either cross platform now or web based applications.

so macs are becoming solely consumer products.
Good bye to professional/specialised software.

 

[Azrael9]

How do you know Apple doesn’t have desktop class ARM chips in the works to compete with Intel? They have full creative license of ARM and can design whatever chip they want... Also ARM server chips are already out there competing with Intel server chips

I'm fully expecting an ARM Mac cpu to kick the Intel cpu right in the nads. I don't think Intel will make the 'ten' count.

Azrael.

 

[jayducharme]

I can use VMWare, Parallels, VirtualBox or even Bootcamp to run and develop for Linux and Windows at nearly native speeds. Moving to ARM would most likely eliminate that option.

Probably not. I remember a nifty little app I used to use with my G5 Mac Pro that allowed me to run Windows. So virtualization software is just that: whatever OS you have, it emulates a different OS. Obviously, Bootcamp will probably disappear (unless Apple has some clever trick up its sleeve). But the virtualization companies will probably do quite well selling new versions of their software.

 

[deconstruct60]

That’s short term thinking. AMD has zero history of sustaining a technical advantage more than a couple of years. They also have zero history of supply the full range of parts needed for an entire lineup of machines.

But if Apple is going to peel off a substantive subset of the entire lineup for their own A-Series derivatives, the question is really whether AMD can cover what is left. As long as it was a super broad, diverse set of CPU products then Intel had leverage.

AMD has gotten off the extremely dubious tactic of trying to go "blow for blow" across the entire Intel CPU product space. If the subset that AMD picks is the area that Apple doesn't want to do then will make them much more viable as a component supplier in that "new boundaries" space.

There is no question AMD could screw things up. For the last couple of years though they are on executing at least as good as Apple is in its subset market targets. Apple just dropped a A12Z to limp along until can get A14X out the door. The gap between A10X and 12X isn't going to much better for 12X to 14X.

AMD has done 14nm -> 12nm -> 7nm on pretty smooth sequence on desktop processors. Looks like Zen3 is a sensible, limited optimization of a narrow set of "could be better" issues at 7nm and will start rolling out at end of 2020. AMD doesn't need a magical technological lead, they need to execute. ( because Intel isn't. )