On Apr 23, 0:57, Tony Duell wrote:
> Indeed. What's worrying is how few people spotted it :-)
>
> When I started in computing all those years ago, I was told 'Don't worry
> about the CPU. It's just a pile of gates and flip-flops like the one's you're
> used to'. Now, admittedly you don't generally see the schematics of a
> microprocesor (although I have understood minis to gate level), but
> there's nothing magic about a CPU.
Agreed. We don't have much on the innards of CPUs at gate level but there's a
1st year course on basic architectures (lots of PDP-11 and -10, IBM, and M68K
stuff), a 2nd year course that teaches about bigger building blocks (pipelines,
ALUs, register banks, cache, etc), and third year stuff on different
architectures (mostly parallel architectures). Everybody has to do the basic
electronic logic course which runs all year in 1st year.
> I hate to say this, but you can't learn this in a couple of practicals.
> Just as you can't learn programming that way. You have to _play_ - build
> circuits, write programs - and keep on at it..
True, but a lot of people just aren't interested. Nevertheless, our first year
course starts with basic gates and has something like 18 or 20 practical
exercises, from investigating glitches in a SPICE model of a NAND gate, to a
traffic light sequencer, a model RAM, a digital die using a PAL, and a few
other things I've forgotten. And everybody has to do it, not just the hardware
buffs.
One of the most popular courses is the 2nd-year follow-up, in which students
are given a problem to solve involving breadboarding a small Z80 system
(CPU/ROM/RAM/LCD/glue + whatever analogue stuff is required). The problem is
different every year, and there's no formal tuition. You can ask the lecturer
or demonstrators any questions you like, and you'll get the answers, but you
have to learn yourself. You don't get taught (in the conventional way).
--
Pete Peter Turnbull
Dept. of Computer Science
University of York
Received on Wed Apr 22 1998 - 20:23:45 BST