Back in the days of floppy disks and cassette tapes and 16--64KB memory spaces it was pretty hard to run into a real performance issue aside from games or sorting.
For the former, converting the inner loops as I described seemed to work, and I suppose it would have at least helped to some degree on the latter.
That said, yes, all the cool stuff was in assembly, and that lasted through to the early GUI days to some degree --- my favourite word-processor was WriteNow for the Mac/NeXT, ~100,000 lines of assembler.
>Sort of?
Back in the days of floppy disks and cassette tapes and 16--64KB memory spaces it was pretty hard to run into a real performance issue aside from games or sorting.
For the former, converting the inner loops as I described seemed to work, and I suppose it would have at least helped to some degree on the latter.
That said, yes, all the cool stuff was in assembly, and that lasted through to the early GUI days to some degree --- my favourite word-processor was WriteNow for the Mac/NeXT, ~100,000 lines of assembler.