Comments by "Lawrence D’Oliveiro" (@lawrencedoliveiro9104) on "Wheeler Jump - Computerphile" video.

  1. DEC’s 12-bit (PDP-5, PDP-8, PDP-12) and 18-bit (PDP-1, 4, 7, 9 etc) machines used this trick. The jump-to-subroutine instruction was called “JMS”, so JMS SUBR would store the address of the following instruction at location SUBR, and transfer control to address SUBR + 1. (cont’d)
    1
  2. (cont’d) By the way, the first prototype of what became Unix was developed on a PDP-7. You can see why they were quick to move to a PDP-11, which had a much nicer architecture with a proper stack.
    1
  3. (cont’d) This would jump to the address stored at SUBR, which was of course the point in the caller at which to resume execution. (cont’d)
    1
  4. (cont’d) There was no return-from-subroutine instruction as such, instead you ended execution of SUBR with an indirect jump: (cont’d)
    1
  5. (cont’d) JuMP I SUBR (cont’d)
    1
  6. And so I just discovered that if I used the right mnemonic for the JuMP instruction, YouTube deletes my comment!
    1
  7. And the latter is a modern superscalar RISC architecture underlying some of the most powerful computers in the world right now.
    1
  8. ... register defined by hardware, the IBM POWER/PowerPC architecture does not.
    1
  9. It’s no big trick to implement a stack in software.
    1
  10. ... had a dedicated stack-pointer ...
    1
  11. ... PDP-11 ...
    1
  12. While the ...
    1
  13. Now it seems I cannot post a comment with “dlo” (backwards) in it.
    1