Skip to content or navigation


Apple Leaving Intel on Macs by 2020—Impact on New Mac Pro

Apple’s leaked ARM Architecture for Mac plans seem exciting but chip transitions tend to hurt deeper more complex applications that are made of multiple software toolsets produced by developers supporting developers.

Advertisement

If Apple abandons the use of Intel CPUs in their future Mac computers they won’t enter a territory they haven’t been before. The tech pioneer used to use IBM and Motorola CPUs since its earliest days. However, in the middle of the last decade, Apple migrated to Intel CPUs and essentially ended its days of competing with Intel PCs on the basis of processor speed.

Welcome Kalamata

Now those days of Apple utilizing a unique CPU architecture that competes with Intel may soon be back, reports a Bloomberg story.

The initiative, code-named Kalamata, is apparently early in its development, which contradicts the intention of making such a move possible by 2020. When Apple moved from PowerPC to Intel the plan apparently took five years. Without getting into the impacts of this Kalamata move—from Intel x86 to Apple-designed CPUs—one of the most confusing questions that emerge from this story involves the upcoming modular Mac Pro.

Modular Mac Pro Woes

As associate editor, Akiko Ashley said the other night on a post-Nvidia GTC event phone call, “why purchase a new Mac Pro next year if in 2020 Apple moves to a new CPU architecture?”

“Exactly!” I quickly replied.

If Bloomberg’s story is spot on Apple will begin migrating Macs to new (possibly A series) Apple CPUs as early as 21 months from now. That, according to the Bloomberg story, will come in stages.

So let’s say the new modular Mac Pro in 2019 ships with Xeons from Intel. At what part of, and how long away, will the next model Mac Pro ship with an Apple CPU?

Bigger Problems Exists

This is just the first troubling aspect of this news. The larger issue has to do with the setback to Apple’s Mac developers. This especially hits hard in the pro apps space like in the CAD and 3D markets that Architosh covers.

Very few industrial heavy-weights in the AEC and manufacturing CAD and engineering software space write every single line of code in their tools. Instead, they rely on third-party software developers who write specialized toolsets, APIs, modeling and visualization engines, and data converter applications. A modern CAD application has a lot of stuff in it made by other people.

01 – For Apple to be successful with pro hardware for Mac users, on a CPU architecture other than Intel x86, it will need to robustly support developers during a long and difficult transition. Ultimately, in the short term, even if done extremely well, there will be a decrease in available CAD and 3D apps for the future ARM-based Macs, not an increase.

When Apple moved to Intel, suddenly Mac developers had possibilities that existing third-party software tools makers could quickly modify their developer toolkits to the new x86-based macOS. That’s largely what happened. Now going the other way will undo that progress and pose problems.

The Good News

The good news is that moving Macs to essentially the same CPU architecture as iPads and iPhones is a move to a very large platform. The Apple iOS platform is larger than the macOS platform. Many core technologies—like the graphics API OpenGL—exists already for ARM-based chips architectures in the form of OpenGL-ES.

Key tools like Siemens’ Parasolid modeling engine, the most popular such engine in the CAD market, is fully written for iOS and is essentially the same. (see: Architosh, “Why Shapr3D Changes the iPad Pro Story From Toy to Tool,” 1 Mar 2018)

Istvan Csanady, CEO and co-founder of Shapr3D, a ground-breaking new MCAD tool for just the Apple iPad Pro with Pencil (it requires the Pencil) told Architosh that the iOS as an operating system and the iPad Pro hardware is not a limiting factor in what Shapr3D can do. In other words, Apple’s iOS platform (software and hardware) appears up to the task on many levels to do everything that big workstation iron has done.

But is it really?

Many leading developers, including Autodesk, have praised the performances of Apple’s iPad Pro hardware and software. But for Apple to successfully move the macOS world over to the iOS (ARM-based chip) world of Apple A series processors, they will need to do many things really well, including helping not just major developers but the smaller developers that those larger developers depend on, to essentially rewrite their x86 legacy code bases over to ARM.

SaveSave

SaveSave

SaveSave

SaveSave

SaveSave

SaveSave

SaveSave

Reader Comments

  1. I was unsure of the PowerPC to Intel switch – and there were some growing pains working with Rosetta not to mention some applications that were abandoned and never ported to Intel. But I have to admit that the move had its benefits. Although I’m not an AutoCAD fan, would they have ever brought it to the PowerPC platform?

    There has been various discussions of merging macOS and iOS for some time now. I can’t say I’m a fan of that kind of thinking. What I want out of a mobile phone or tablet is much different from what I want out of a desktop machine – much of the reason I’m disappointed by Apple’s lack of attention to the Mac Pro line. I’m still using a 12-core 2010 Mac Pro “Hackintosh” with dual NVidia cards and an external secondary PSU to support VRay. Moving to ARM chips would obviously be a move in this direction. Will it leave the power users whose workflow does not “fit” on tablets and laptops with no options with Apple?

  2. Thomas,
    These are all really excellent comments and points. Very interesting to hear about your 12-core 2010 Mac Pro hack that powers your workflow to this day. I think you and I have spoken about Apple’s Mac Pro woes at depth before. I will share a couple of thoughts here that I think is important.

    Awhile back on Architosh when Apple announced their true commitment to a new Mac Pro, I mentioned the new machine as being useful also for emerging power-hungry applications like AI and machine learning. This was prescient on an accidental level; I think Apple wants to do a lot with data, machine learning and AI and they want lots of security and control in that domain. To control their security destiny going custom CPUs seems to offer advantages.

    Beyond the importance of AI and machine learning, and in particular its importance to Apple’s third-party developer community, the modular Mac Pro folks like you and me are dreaming about must be performant within the rival Intel community. It must, ideally, out-perform Xeons and i9’s. And an architecture transition must go at least equal or better than Rosetta. Have you read Tech Crunch’s article here below?

    Apple’s 2019 Mac Pro will be shaped by workflows?
    https://techcrunch.com/2018/04/05/apples-2019-imac-pro-will-be-shaped-by-workflows/

  3. I did read the TechCrunch article and that they’re talking about actual workflows is somewhat encouraging. There is that element of Apple telling users what they need – you don’t need disk drives, don’t need optical drives, etc. In a lot of cases they’re just at the forefront of the curve, but then in other things they miss. Sorry, a laptop or an iMac just doesn’t meet MY needs. At the very least I would have to tie it to a bunch of external hardware that I DO need instead of just being able to put it into my Mac Pro box.

  4. Thomas,
    Something tells me this is going to be different. I also think edge compute will become incredibly important going forward, the many uses of AI in various industries and in the field. So performance again matters and Apple to be relevant in the AI world needs to catch up.

  5. Something else to consider here is that the number of developers on the Mac platform is dwindling compared to iOS. There are so many apps that have been abandoned or barely see updates. Initiatives like Marzipan and Kalamata will naturally bring developers back to the Mac. Of course there are many problems with changing chip architectures, but there are good things too. We’ll see if it happens. The article title makes it sound like it’s a done deal but I have reservations.

  6. Evan,
    It is clearly not a done deal and Bloomberg’s report noted that things like this could be abandoned. The 2020 timeline would strongly suggest they are near or beyond the point of no return on this chip decision. Whether all Macs will move quickly to the new architecture or not is another aspect that speaks possibly to the 2019 Mac Pro.

    I have not be tracking Mac developer counts. Are the numbers really declining, where do you see evidence of that? My view from the MCAD industry is that macOS is growing, not declining.

  7. I recognize the volume inherent in the iOS platform, but I also recognize that although many people may only need a small commuter car, there will always be the need for pickup trucks and larger – just as Steve Jobs is quoted saying in 2010 (http://www.businessinsider.com/steve-jobs-quote-apple-computers-metaphor-2017-4). Tim Cook makes some similar points as well (https://www.macrumors.com/2018/04/19/tim-cook-still-opposed-to-merging-mac-ipad/).

    Case in point: I’ve watched as iWork features were eliminated as they combined the macOS and iOS versions and am reminded of the office I worked in that tried to use AutoCAD and MicroStation interchangeably. Rather than using the particular features of one or the other as a benefit to a particular project, they tried to condense everything down to the lowest common denominator – features available in only one package were not used. At the time, MicroStation used 64 numbered layers, so we were not allowed to name layers (or use more) in AutoCAD. The power users for each platform were constantly frustrated.

    iOS has come a long way and offers many advantages. However, my workflow involves an ergonomic keyboard, 3D mouse, and Logitech G13 gameboard for input; multiple screens; and there are still times I walk away from my Mac Pro workstation to let it crunch through a render or computation – despite its 3.24Ghz 12-core processors and dual GPUs.

    Perhaps the ARM chips can support this level of work, but I’m concerned that needs of power desktop users on one side vs. mobile users on the other would draw these chips in different directions. There are already issues with Apple’s form factors (Macbook/iMac) dictating lesser performance to achieve smaller, slimmer packages.

Leave a Reply