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

[throAU]

Try VMWare Fusion 15. Metal support, DX11 support. Works flawlessly here, and its a 2013 MBP 13" Late i7.

You realise the CPUs in the current iPhone are more powerful than your 13" MacBook Pro yes?
[automerge]1587701762[/automerge]

Then what will be the main issue with ARM based Mac if the compatibility is not an issue? Performance?

Running legacy code that people refuse to upgrade because "they paid for it" in 2009 or whatever.

This is why much software is going to subscriptions. Because continued platform support (even having to re-write/update for new OS on the SAME CHIP), and security fixes for the modern connected world demand it. You can't just release a package and wipe your hands of it anymore, fixes need continual funding.

This is why the argument that Adobe (for example) will never port their stuff to ARM is laughable

• they already did port much of it to ARM for the iPad versions
• they already need to do constant recompiles and re-releases of their x86 software to fix bugs. recompiling for ARM is something they are likely already doing for much of their code already - the UI specific stuff is handled by macOS libraries anyway.

Yes, it will mean VMware fusion either goes away, or somebody releases an emulator for x86. QEMU based probably the technology to do this stuff already exists, it just depends how fast you want it to run. If you NEED VMs then you will maybe be out of luck. The vast majority of users do not, they need APPLICATIONS to work.

Apple are fortunate that the majority of their apps have most of the "hard work" done by system libraries which aren't even part of the application. i.e., the App makes calls to macOS library, which does the work. Those libraries will be native arm, and only the glue code that holds it all together will need to run in emulation.

This is how x86 was just as fast at running PPC native apps when or shortly after Apple shifted (to x86) via Rosetta, and how performance on x86 running PPC software outperformed native PPC hardware in a lot of cases within 1-2 generations (Never mind how much faster x86 was by then).

Same thing will happen with ARM. Apple's software architecture enables this.

 

[MikeZTM]

Then what will be the main issue with ARM based Mac if the compatibility is not an issue? Performance?

Surprisingly it's cost.

Tapping a new silicon die isn't cheap.
Even if a new MacBook with ARM may run 2x as fast as current Intel chip and last 2x battery life. It is still cheaper to purchase mass produced chips from Intel than design and fab their own.

iPhone chips are sold in dozens of millions by just one SKU each year.
MacBook have multiple SKU for just one size (13/15/16 with i5/i7) and are sold in less than a million for each SKU.

If Apple finally move to ARM it just mean they are really care about customer instead of pure margins.

 

[endlessike]

It's funny because people said this exact thing in 2004-2005 in the run-up to the Intel switch. People said, developers won't update their software, the Mac is too small of a market.

They said it will make people think of Macs just like PC's, less special and they would just buy PC's. They said it would cause piracy of Mac OS X and result in people buying Dells and loading Mac OS X on them (the fear here was duplicating what happened with Mac OS 8 when Apple licensed it to vendors and no one bought Apples hardware which was one of the many reasons Apple almost went bankrupt).

But you know what happened? People saw the value in a higher performing computer with better battery life. Developers followed their customers and released updated versions of their software to support the new architecture and Apple in the interim provided Rosetta Stone so we could run the Apps that didn't get an x86 recompile.

Fact is if these computers offer better battery life or higher performance (and it looks like they'll offer both based on the big.LITTLE architecture of the rumoured 12 core chip) it will do great in the market and developers will jump on board.

I don’t really remember anything of the sort. I remember almost everyone being very much in favor of the switch away from PPC because it had fallen so obviously far behind x86, and people were extremely excited about the prospect of something like BootCamp.

Hard for me to see how they are going to sell users on the loss of Bootcamp, or some kind of Windows on ARM Bootcamp. A lot of people see Bootcamp as an absolutely killer feature. I don't see my Steam library running on Windows ARM.

 

[Nugget]

Running legacy code that people refuse to upgrade because "they paid for it" in 2009 or whatever.

I'd argue that losing virtualization is a legitimate issue as well, and can't be "victim shamed" like you're doing here. Losing VMware, Parallels, and Docker Desktop will be a fatal blow for many of us.

 

[MikeZTM]

You realise the CPUs in the current iPhone are more powerful than your 13" MacBook Pro yes?
[automerge]1587701762[/automerge]


Running legacy code that people refuse to upgrade because "they paid for it" in 2009 or whatever.

This is why much software is going to subscriptions. Because continued platform support (even having to re-write/update for new OS on the SAME CHIP), and security fixes for the modern connected world demand it. You can't just release a package and wipe your hands of it anymore, fixes need continual funding.

Yes, it will mean VMware fusion either goes away, or somebody releases an emulator for x86. QEMU based probably the technology to do this stuff already exists, it just depends how fast you want it to run. If you NEED VMs then you will maybe be out of luck. The vast majority of users do not, they need APPLICATIONS to work.

Apple are fortunate that the majority of their apps have most of the "hard work" done by system libraries which aren't even part of the application. i.e., the App makes calls to macOS library, which does the work. Those libraries will be native arm, and only the glue code that holds it all together will need to run in emulation.

This is how x86 was just as fast at running PPC native apps when or shortly after Apple shifted via Rosetta, and how performance on x86 running PPC outperformed native PPC hardware in a lot of cases within 1-2 generations.

Same thing will happen with ARM. Apple's software architecture enables this.

VM already works on ARM. In fact all current Android device is running under a hypervisor.
VMware could still exist but only runs ARM OS.

Maybe bootcamp will still exist but running Windows 10 ARM.

 

[throAU]

I'd argue that losing virtualization is a legitimate issue as well, and can't be "victim shamed" like you're doing here. Losing VMware, Parallels, and Docker Desktop will be a fatal blow for many of us.

Yup, and I'm a VMware Fusion user as well.

However it is increasingly less relevant. More server stuff is moving to the cloud which you can administer from anything with a browser, more apps are cross platform, and more of my stuff is shifting to open source platforms such as Linux which is available to run on ARM anyway.

ARM natively supports virtualising ARM. You can already likely emulate x86 on ARM via QEMU. Apple will no doubt provide a binary translator like Rosetta as well.

yeah, there will be people who lose out, but the vast, vast majority of Apple's customers will see better performance, improved battery life, and Apple as a company will have more control over their roadmap. They'll be able to design processors for THEIR application, not use general purpose parts from intel that may or may not line up with what they want to achieve.

This shift makes total sense for both APPLE and like 95% plus of their customer base. Additionally, as above it gives them total control of their product roadmap, rather than stagnating for a decade because intel CPU development fell in a 10nm hole.

 

[cmaier]

Surprisingly it's cost.

Tapping a new silicon die isn't cheap.
Even if a new MacBook with ARM may run 2x as fast as current Intel chip and last 2x battery life. It is still cheaper to purchase mass produced chips from Intel than design and fab their own.

No, it most certainly is not.

 

[Nugget]

However it is increasingly less relevant. More server stuff is moving to the cloud which you can administer from anything with a browser, more apps are cross platform, and more of my stuff is shifting to open source platforms such as Linux which is available to run on ARM anyway.

For me "the cloud" means "x86 docker containers" and the day I can't run and produce those is the day I have to begrudgingly move to a Linux machine on x86 hardware.

 

[high heaven]

Surprisingly it's cost.

Tapping a new silicon die isn't cheap.
Even if a new MacBook with ARM may run 2x as fast as current Intel chip and last 2x battery life. It is still cheaper to purchase mass produced chips from Intel than design and fab their own.

iPhone chips are sold in dozens of millions by just one SKU each year.
MacBook have multiple SKU for just one size (13/15/16 with i5/i7) and are sold in less than a million for each SKU.

If Apple finally move to ARM it just mean they are really care about customer instead of pure margins.

I still not sure if it's really easy to port because when Apple announced Catalina, they ditched 32 bit and still having a lot of bugs and errors.

 

[Emanuel Rodriguez]

I'd argue that losing virtualization is a legitimate issue as well, and can't be "victim shamed" like you're doing here. Losing VMware, Parallels, and Docker Desktop will be a fatal blow for many of us.

All popular operating systems (even many you probably haven't heard of) except macOS have ARM versions at this point. I find it likely that virtualization software will be ported to the new macOS on ARM, and we'll be able to install those operating systems, and run them as we did on Intel (ARM supports virtualization extensions similarly to how Intel does). There will be a transitionary phase, but eventually, all software will be compiled for ARM, and run on ARM-based operating systems.

Old software which won't be updated is a concern, but that software will continue to run on old hardware anyway, and at least on macOS, will continue to run as long as Apple provides the expected emulation layer, which will presumably be removed in a future OS update, but users won't be required to update to that (similar to the PowerPC emulation in Rosetta, which was removed in Snow Leopard).

A lot of future-looking here, but I've seen Apple's transitions, and they usually work more or less the same way.

 

[Nugget]

All popular operating systems (even many you probably haven't heard of) except macOS have ARM versions at this point. I find it likely that virtualization software will be ported to the new macOS on ARM, and we'll be able to install those operating systems, and run them as we did on Intel

People use these technologies to run applications, not to run operating systems. Intel is where the applications are, and sadly Apple have no influence to change that reality. This is going to be tragic for anyone whose workflow involves multiple platforms. I do agree with others -- we are likely a small percentage of the Apple user base. Apple can probably afford to lose us. But please just let me mourn. If this happens it will be the end of a 30 year customer relationship I've had with Apple, and that's sad and scary to me. Linux has come a long way, but it's no macOS. Windows has come a long way lately, but it's no macOS. It's going to be a bumpy and uncomfortable transition for me.

 

[newyorksole]

Their recent mistakes and after Catalina dropped 32 bit support, even my dad is ready to leave that ship to return to Windows. One of his favorite apps TextWrangler, is only 32 bit and will not be upgraded. He does not want to use their successor (BBEdit) and is freaking furious at how Apple keeps changing the things in the OS all the time. Plus, his camera surveillance apps for our home does not work on mac either, so he is ready to leave the ship. I'll be ready too, if Macs go only ARM based.

That definitely is frustrating. I’d probably feel the same way if I was in the same boat. We’ll see what happens. I did read a lot of complaints that had to do with Catalina and the constant changing OS.

These next couple years are going to be very transformative.

 

[Emanuel Rodriguez]

People use these technologies to run applications, not to run operating systems. Intel is where the applications are, and sadly Apple have no influence to change that reality. This is going to be tragic for anyone whose workflow involves multiple platforms. I do agree with others -- we are likely a small percentage of the Apple user base. Apple can probably afford to lose us. But please just let me mourn. If this happens it will be the end of a 30 year customer relationship I've had with Apple, and that's sad and scary to me. Linux has come a long way, but it's no macOS. Windows has come a long way lately, but it's no macOS. It's going to be a bumpy and uncomfortable transition for me.

No, virtualization software runs operating systems.

My point is that I would expect that most software will eventually be ported to those ARM-based operating systems, some of which are already well-established (Linux on ARM for example). There will be a transitionary phase, but I do believe it'll all eventually be sorted out.

 

[Nugget]

No, virtualization software runs operating systems.

My point is that I would expect that most software will eventually be ported to those ARM-based operating systems, some of which are already well-established (Linux on ARM for example). There will be a transitionary phase, but I do believe it'll all eventually be sorted out.

Yes, but nobody on earth cares that they can run Windows for ARM. That's just a means to an end -- running an application they want to be able to run. So make a list of all those wonderful Windows for ARM applications that people are dying to run. I'll wait...

That was my point.

I need to run and create applications and those applications run on x86. I can't afford to wait years for all the other major operating systems to embrace ARM (if that even happens at all). I need to do my job right now.

 

[cmaier]

Yes, but nobody on earth cares that they can run Windows for ARM. That's just a means to an end -- running an application they want to be able to run. So make a list of all those wonderful Windows for ARM applications that people are dying to run. I'll wait...

That was my point.

I need to run and create applications and those applications run on x86. I can't afford to wait years for all the other major operating systems to embrace ARM (if that even happens at all). I need to do my job right now.

“My applications need OS/360 and run on COBOL!”

 

[Emanuel Rodriguez]

Yes, but nobody on earth cares that they can run Windows for ARM. That's just a means to an end -- running an application they want to be able to run. So make a list of all those wonderful Windows for ARM applications that people are dying to run. I'll wait...

That was my point.

I need to run and create applications and those applications run on x86. I can't afford to wait years for all the other major operating systems to embrace ARM (if that even happens at all). I need to do my job right now.

Which is why I'm being smart and buying the 14" MacBook Pro Apple releases this year so that I can ride out the transition. By the time everything settles down, I'll be about ready to upgrade.

Really, there's rarely a reason to panic in life.

 

[Nugget]

“My applications need OS/360 and run on COBOL!”

I see the humor, but in fact that was true at one time (Well, maybe the debate was about upgrading to some new, fancy ANSI dumb terminal when the System/360 only worked with EBCDIC terminals...) For me, right now, it's true for x86. What can I say? I'm not really alone in this position either. Our production systems are all Kubernetes x86 docker containers. Bazillions of them. That's not changing any time soon, and I need to be able to locally produce, debug, run, and utilize those containers. In macOS currently this is a really effective workflow, with Docker Desktop for Mac natively supporting docker pulling and pushing all day long.

macOS on ARM means I can't do that. At least not in any sane way without introducing performance-sucking emulation and introducing uncertainty as to whether or not my local tests are actually valid and informative about how the code will behave in production.
[automerge]1587704728[/automerge]

Really, there's rarely a reason to panic in life.

I'm not panicked, I'm just sad.

 

[PickUrPoison]

I am pretty sure these 1st-gen ARM-based laptops are going to be pricey, right?

Why would they be?

Are Apple’s ARM chips more expensive than Intel’s CPUs? (Spoiler: they are not. For example, Apple’s A13 is estimated to cost $65.)

 

[throAU]

That definitely is frustrating. I’d probably feel the same way if I was in the same boat. We’ll see what happens. I did read a lot of complaints that had to do with Catalina and the constant changing OS.

These next couple years are going to be very transformative.

If a a guy is unwilling to pay (something like $60) for the paid version of a text editor that is his core application that he runs on his $x,000 computer, and is willing to change platform over it then he's a lunatic.

Normal people would either pay for the upgrade to Bbedit (I did, I'd used Textwrangler occasionally for years) or switch to one of the many free text editors that do still run on MacOS.
[automerge]1587706472[/automerge]

 

[PickUrPoison]

What I mean is, is Apple going to be charging a premium for these laptops, as compared to their Intel-based counterparts? In other words, will I have to pay around $2000 for the 11-inch/13-inch variants, as opposed to the 13-inch Intel-based MacBook Pro, which starts at $1299?

Don’t be frightened. There’s no reason to think prices will increase 50% as a result of lower costs.

 

[im_to_hyper]

I really like your thought process. I'm almost wondering if they won't switch to a semi-custom chip in partnership with AMD.... a hybrid laptop, perhaps? 8 AMD cores and 4 Apple Cores on the same die. The OS and Apple apps use the low-powered ARM CPU for basic... OSy things... so that the x86 Apps can run even faster without any hindrance.

 

[throAU]

I really like your thought process. I'm almost wondering if they won't switch to a semi-custom chip in partnership with AMD.... a hybrid laptop, perhaps? 8 AMD cores and 4 Apple Cores on the same die. The OS and Apple apps use the low-powered ARM CPU for basic... OSy things... so that the x86 Apps can run even faster without any hindrance.

Battery life, heat, board space and the fact that for 99.99% of their userbase x86 vs. arm is irrelevant so long as the apps work.

Doing that won't make a clean break from x86 and leave them with the problem they already have - dependence on a third party for their product roadmap.

Holding onto legacy stuff is why windows has been such a mess for decades. It adds complexity, bloat and more opportunity for bugs and security issues.

Its time to let go...

 

[im_to_hyper]

Its time to let go...

Queue the Frozen song...

 

[PickUrPoison]

ram will still be like 400 dollars for 16gbs

Why would switching the CPU architecture cause RAM prices to double?

 

[Janichsan]

That can't happen. Remember that Apple needs Macs so professionals can create the apps and content consumed on iOS, iPadOS, tvOS, and watchOS.

Sure about that?