13 Apr
2002
13 Apr
'02
11:08 a.m.
I'd say they're pretty practical for shallow stacks. However, when one is
exploring the realm of stack-driven architectures, you need a pretty deep one.
Admittedly, I completely spaced the shift register approach to stacks, but
that was because the deepest stack I'd ever seen with shift registers was
about 2K deep. That was an old-timer, though. It used a bunch of 2513 shift
registers (in TO-5 cans) on a board dedicated to that purpose.
see below, plz.
Dick
----- Original Message -----
From: "Ben Franchuk" <bfranchuk(a)jetnet.ab.ca>
To: <classiccmp(a)classiccmp.org>
Sent: Saturday, April 13, 2002 9:15 AM
Subject: Re: TTL computing
That seems reasonable, but in what time-frame are you thinking?
Richard Erlacher wrote:
(computer science stuff)
subroutine
I told it was a semantic quagmire.
I
really don't care what it was called.
> > Of course you can build a stack without a counter. In the real world a
> > shift register will do it (I've used a row of 'F194s to make a
> > stack). Or a pile of bits of paper with
numbers written on them :-). You
> > do not have to make a stack (particulary not in the 'theoretical
world')
> > using a counter
> >
> Oh forgive me for trying to be practical. It takes a LOT of shift
registers
to build what
you can build with an up/down counter and a RAM.
Shift register stacks do have the advantage of being fast. Did not the
8008 or the 4004 use a 8 level stack for subroutine calls. A calculator
chip at that time only needed 4 functions.
--
Ben Franchuk - Dawn * 12/24 bit cpu *
www.jetnet.ab.ca/users/bfranchuk/index.html