cctech digest, Vol 1 #370 - 5 msgs

From: Paul Shubel <paul.shubel_at_verizon.net>
Date: Sat Mar 8 16:15:01 2003

> I tend to agree, but I would specialize that even further. As time goes
on,
> I become more convinced that ALU width should be the determinant of
bittedness.
> The 6809 requires multiple ALU passes for a 16 bit operation because the
ALU
> is 8 bits wide. The 68000 is a 16 bit processor for the same reason.
>
Hi Everyone,

I have thought about this topic at length...............
This is a much debated issue because microprocessors tend to have
8 and 16-bit instructions, while microcontrollers tend to have instructions
that are limited to 8-bit operands. Really, does it matter what size the
ALU is in silicon within the chip? I say no! When you are coding in
assembly
you're interested in what instructions are available to you.

For example, the Z8, 8051, 6805, PIC16, etc.. have mainly 8-bit operations
with few if any 16-bit operations. Note, please leave 16-bit jumps out of
this
discussion. The 6502 uses 8-bit operands also.

On the other hand, the 8085, Z80, and 6801 are considered 8-bitters in the
media
but they have 16-bit operations that make them more powerful than
the example micro-controllers above. Side note: Since these processors
can handle 16-bit operands with ease, pointer arithmatic is very compact.
Pointer arithmatic is very important when porting to high-level languages!

I consider the 68HC11/6809 complete 16-bitters. Why not! They both have
oodles of 16-bit operations, like the ever-useful 16-bit compare immediate!

So here is how they stack-up:

8-Bitters: 6502, 6805, Z8, 8051, PIC16, TMS370, ST6

8/16-Bitters: 8085, Z80, ST7

16-Bitters: 68HC11, 6809, 8088/8086, 65816, 80C166, ST9

16/32-Bitters: 68000, Z8000, 80286

32-Bitters: 68020, 80386/Pentium, ARM, MIPS, NS32000

Cheers,

Paul
Received on Sat Mar 08 2003 - 16:15:01 GMT