Re: FPGApple + ARM?

["Michael J. Mahon" <mjmahon@aol.com>]

sicklittlemonkey wrote:
> You're raining on my parade, man. Talking like a compiler writer!
> Hopefully there's another "Not that I don't love it, you understand!"
> coming! ;-)

As a computer architect, I had to talk like: 1) a compiler writer,
2) an OS designer, 3) a logic designer, and 4) a chip designer.  Only
by weighing the costs and benefits across the whole spectrum (taking
trends into account) can you get a "sweet" design with a reasonably long
effective lifetime.  This kind of design is done by a great team
(*not* to be confused with a committee!).

And I loved it--but I'm retired now.  ;-)

Of course, as we all know, "successful" architectures are made in the
marketplace, not the R&D lab...  ;-(

> Anyway, from the points you make (all valid of course) the nature of
> RISC tends towards a division into processing instructions and memory
> access instructions. Combined with a relatively large set of general
> purpose registers, this leads to the kind of orthogonality I was
> referring too. In the ARM any operand can be "barrel-shifted" for
> "free" - all within one cycle for instance.
>
> As you say, not much help there from 6502 experience. I was quite taken
> aback when I discovered the ARM had no stack pointer!

Yep, that's what general registers are for!  ;-)

-michael

Music synthesis for 8-bit Apple II's!
Home page:  http://members.aol.com/MJMahon/

"The wastebasket is our most important design
tool--and it is seriously underused."