Re: life extension of old computers thanks to opensource Contiki

["Peter de Vroomen" <peterv[at][spamblock]jaytown[dot]com>]

> > > > value in learning the M68K instruction set because it's practically
a
> > > > dead family.
> > >
> > > Sadly the instruction set is a dead family.
> >
> > I've been working with Coldfire a lot of late.  Using assembler on it is
like a
>
> Since Palm went to ARM, ColdFire is more or less a dead end, IMHO.

What I liked about the M68K instruction set is the way it still was tailored
to handcoded assembly programmers. The Commodore Amiga became a success
because all the C64 6510 hand coders got a processor which was easier to
code for, was faster, and more versatile too.

The 8088 on the other hand had such an ugly instruction set that it was
shunned for years by the people who matter :D. Thats the main reason why the
PC had so few games in the 80's. Only with the 80486 Intel's instruction set
was cleaned up so much that I would dare to hand-code assembly for it. And
that really means I will *never* hand-code assembly programs for it *ever*,
as today's programming tools are so much better and the processors are so
much faster.

The ARM is a RISC processor, and RISC processors are not really suitable to
hand-coding. Assembly code for a RISC processor needs to be optimized to
have it make use of all the advanced features. Unoptimized code is often
slower than unoptimized code of CISC processor, while optimized code may run
MUCH faster than optimized code of a CISC processor.

So stepping over to ARM and hand-coding for it to learn about assembly
language is not very efficient. Better to start on the M68K and then going
on to ARM. Things like 'pipeline' and 'branch prediction' will be a lot
easier to understand for a student.

Did you know, by the way, that the 6502 processor (or at least the 65xx
family) is the worlds first microprocessor to have a pipeline (or so I was
told at school)?

PeterV