2020 may finally usher in ARM-based Macs

Looking ahead to 2020, Macworld’s Dan Moren looks at some of the big things to expect (and not) out of Apple in the coming year.

Dan Moren, Macworld:

Apple announced its transition from PowerPC to Intel processors in June 2005, just over ten years after it had made the jump to PowerPC from the Motorola 68000 architecture that had powered every Mac in the decade before. Fourteen years since that last transition, it’s starting to seem like we’re overdue for the next jump…

Recent chips powering the iPad and iPhone have proved to be screamers, in some cases outperforming Macs in benchmark tests. These chips also tend to consume less power than the current processors in Macs, helping provide better battery life. Not to mention every non-Mac device that Apple builds—from the Apple Watch to the HomePod to the Apple TV to those iOS devices—uses processors that Apple itself designs. That leaves the Mac as the odd man out in the company’s lineup…

This seems more like a “when” question than an “if” question. And the smart money suggests that 2020 might be the year that Apple finally makes this rumor a reality.

MacDailyNews Take: WWDC 2020 is going to be quite the show!

21 Comments

    1. PowerPC didn’t fail because RISC was inferior to CISC. It failed because Intel (along with Microsoft) commanded overwhelming duopoly market power, and because Intel had so damned much money, with so much investment riding on x86’s success, that it could have built a Turing Machine out of chicken bones that would outperform the products of the relatively underfunded (“beleaguered” even) AIM consortium.

      Given an even playing field, RISC would have kicked CISC’s backward ass into the 19th century. And now Apple has the resources and incentive to prove it.

      Watch this space.

      1. More on that:

        RISC is vastly superior in terms of cycles per watt. That is the metric that matters in a battery-powered world. It’s not relevant that G5 was inefficient; that was the last gasp of Motorola’s dying chip-building effort. Had Motorola had the funds (that is, if Macintosh had had the market power to afford incentive for real investment into chip technology), well-known techniques of pipeline optimization would have crushed Intel’s x86.

        CISC’s primary advantage over RISC is bytes per instruction. (RISC often requires two or three simple instructions to do what CISC can do in one big slower operation.) That was important back when RAM was expensive and memory cache sizes were small. None of that is relevant now.

        (Matter of fact, even Intel knows this; that’s why it made a valiant attempt to make Atom work. It was basically an incompetent grasp at the efficiencies of RISC.)

        These days the smart money is on ARM.

        1. RISC gets more cycles per watt because it does less per cycle. Intel’s CISC architecture queues up instructions such that they operate in parallel. It is not uncommon for RISC architectures to take 3x the cycles of Intel CISC architecture to perform the same task.

          1. There’s nothing inherent about CISC that allows it to operate in parallel more effectively than RISC. The main thing that makes CISC pipelines deeper is a fat R&D budget. And now Apple has that. I’m betting on Apple on this one.

            1. Not inherent but, for Intel, it is there. RISC processors don’t do this. I work day in and day out on software that runs on both Intel and ARM. I’m seeing a 3x difference in cycle counts as a general rule.
              The gap gets much larger if you can take advantage of Intel’s AVX2 or AVX-512 instructions which operate on 256-bits or 512-bits, respectively. No ARM even approaches this.
              ARM is good for low power but, for real processing power, I have zero interest in an ARM.

  1. Apple blew when they moved to Intel and didn’t go with Sun Sparc chips. Sun was not in good shape at the time just before Oracle acquired them. Apple could of bought the Sparc business from Sun at a good price and had their own RISC chip in 2005 and by now have many years of porting OS X and Apple software to RISC. Apple would of had everything in one purchase chip designers and manufacturing in one buy. But no they went Intel like everyone else and Oracle bought Sun and they are heart of Oracle servers and other hardware.

    Apple’s chip have been good for small devices on a lightweight OS, but how long before they ramp up to running a full size computer, OS, and pro level applications like Logic and Final Cut.

    1. Newsflash. Chips is iOS devices are RISC chips. They excel at simple single tasks. They start to fall apart under heavy multi process usage. CISC chips (intel or AMD) are better suited for workstations.

    1. The G5 was a dead end because Motorola and IBM didn’t think smaller faster chips were the future (THEY WERE WRONG), Intel asked by the best CEO in tech to design a cpu chip for the iPhone said no (Intel is now a large tech boat without a rudder), and all of those American companies had a chance to be big profitable part of the future and they all said no to Steve Jobs, Apple will leave Intel just a matter of time. ALL OF THOSE COMPANIES FORCED APPLE TO DESIGN THEIR OWN CPU’S CHIPS. My portfolio thanks you.

        1. Steve asked each of those companies (IBM, Motorola, and Intel) to step up and make smaller and faster CPU’s for laptops and mobile devices, they did not (they all said no), Motorola is dead and Intel and IBM are on the long slow path to irrelevance (joining Xerox, and HP). See the declining profit and revenues. The size of the G5 means nothing. Apple at the time (since 2001) was working on the iPad and iPhone…..

    2. The G5 might have been a dead end for laptops and small desktop computers because of its power consumption and heat dissipation issues, but the Power Instruction Architecture series of CPUs wasn’t. IBM continued to develop them and still makes them fifteen years later. It has used them in most of its minicomputers and similar devices during all that time. There are at least two Power-powered supercomputers due to come on line in 2020. The processors are widely used in open computing environments.

      Apple made the right move when it went Intel because those processors were getting faster without getting ridiculously hotter. The ability to run Windows software without processor emulation was another huge plus. It was easier to run RISC emulation on a CISC processor than vice-versa, so the faster Intel Macs could run existing PowerPC Mac software without taking a huge performance hit. Unless the A-Series (or whatever the iPhone-ish processor for Macs is called) is a LOT faster than the Intels it is replacing, emulation will be painfully slow.

      Not just Windows, either. All existing Mac software is Intel code as well, and will be until the developers go back and recompile their source code into “fat binaries” or whatever they will be called (and that assumes that source-code tweaks won’t be needed). Some of the software that we use in our daily workflows may never be updated, or only after years of delay. It will not be a painless process.

  2. The dead end for IBM chips is IBM, their earnings, revenue and profit has been on downward spiral since 2012, Apple’s move to Intel was right for the time but that time is coming to an end, the need to combine their OS and hardware together will cut Intel out.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.