Re: Continuations

On Tue, 16 Jul 2024 16:00:59 +0000, John Savard wrote:

On Mon, 15 Jul 2024 01:59:58 -0000 (UTC), Lawrence D'Oliveiro
<ldo@nz.invalid> wrote:
>

It only has to be Turing-equivalent (or Lisp-equivalent) to be a
“general-
purpose” machine. There’s nothing in that criterion that requires gotos.

>
It's true that Turing-equivalence is sufficient to allow a computer
with sufficient storage available to do anything.

But we also want that stuff to be fast (under some metric).

>
However, the term "general-purpose" can have other meanings. The sense
in which I was using it was: a machine that is _efficient_ at running
programs written in all popular programming languages for the kinds of
applications for which computers are normally used. Thus, the IBM
System/360 was general-purpose since it had string manipulation
instructions and decimal arithmetic in addition to floating-point and
binary integer arithmetic.

I think that history shows that those string facilities were, at best,
overkill. None of the RISC machines had any of that and were faster
and at least as general purpose as 360.
Especially the COBOL stuff like EDIT and EDIT-and-MARK.
On the other hand, I am advocating that some Elementary Functions be
raised to first class citizens of ISA so that SIM() is 20 cycles instead
of 150 cycles. Why, because today, we understand enough about the cal-
culations, polynomials, error sources, to build in HW things that are
just as accurate as SW can build and nearly 7×-10× faster.

>
John Savard