• mihies@kbin.social
    link
    fedilink
    arrow-up
    1
    ·
    6 months ago

    It’s not just node, it’s also the design. If I remember properly, ARM has constant instruction length which helps a lot with caching. Anyway, Apple’s M CPUs are still way better when it comes to perf/power ratio.

    • sugar_in_your_tea@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      1
      ·
      6 months ago

      Yes, it’s certainly more complicated than that, but the lithography is a huge part since they can cram more transistors into a smaller area, which is critical for power savings.

      I highly doubt instruction decoding is a significant factor, but I’d love to be proven wrong. If you know of a good writeup about it, I’d love to read it.

      • mihies@kbin.social
        link
        fedilink
        arrow-up
        2
        ·
        6 months ago

        Here, I’ve found an article. M CPUs are also much more integrated, like a SoC, which gives them further performance/power advantages sacrificing extensibility.