1. 3
  1. 1

    Lots of architectures listed, lots reminders that everything was better than x86 and still is.

    I’m particularly fond of the 68000. Despite the first models having 16bit ALU and bus, and 24bit addressing, the isa was 32bit from the beginning, with 8 data and 8 address registers, supervisor/user modes and a flat view of memory. They had very clear vision of the direction computing was going, and made a reasonably future-proof design.

    Meanwhile, 8086 was 16bit, had no user mode, was register-starved, and memory had to be addressed through 64KB segments, which is very annoying.

    1. 3

      The 68k was forward thinking in some ways, but a bit behind in others. The orthogonal ISA and other CISCisms like indirect addressing created a lot of decode complexity that makes it hard to scale up. (Not helped by Motorola’s own fab issues throughout the 80s and into the early 90s.) Ironically, x86 is a bit better here and avoided a lot of those mistakes.

      Meanwhile, 8086 was 16bit, had no user mode, was register-starved, and memory had to be addressed through 64KB segments, which is very annoying.

      A lot of 808x limitations make sense given context.

      1. 1

        68000 still managed to be typically earlier to milestones, and have higher performance all the way through 68060, which was released earlier than the pentium yet easily had twice the performance per clock.

        Then Motorola abandoned the 68k line to focus on powerpc.

        1. 3

          Then Motorola abandoned the 68k line to focus on powerpc.

          You forgot the failure of 88k. And the major scaling and fabrication difficulties along the way.