Author: Albert Silver
Date: 21:08:56 03/09/00
Go up one level in this thread
On March 09, 2000 at 23:11:22, Robert Hyatt wrote:
>On March 09, 2000 at 22:28:33, Tom Kerrigan wrote:
>
>>On March 09, 2000 at 22:18:12, Eugene Nalimov wrote:
>>
>>>Is 8080 16-bit processor? It has 3 register pairs, each 16-bit wise, and there
>>>were operations that worked at those - addition, load, store. And z80 adds more
>>>16-bit instructions - e.g subtraction with carry.
>>>
>>>On 68000 32-bit operations worked much slower than 16-bit operations, even for
>>>data on registers.
>>>
>>>Eugene
>>
>>How wide was the register bus in the 8080? My guess is 8 bits. I believe the
>>register bus in the 68000 was 32. I don't think instruction speed has a place in
>>determining the "bitiness" of a processor.
>>
>>-Tom
>
>
>I think the easy test is that every N bit processor I have ever used also had
>some 2*N bit instructions. IE when you do a multiply or divide, you have to do
>that or take a big integer precision hit.
>
>My old xerox sigma9 was 32 bit, but you could use several instructions that
>would take an even-odd register pair and treat them as one. Ditto for the old
>32 bit vax. Older PDP-11's had 16 bit everything, but had some 32 bit
>instructions as above...
>
>Simple test is "what did Motorola call this microprocessor?" 16 bit or 32 bit?
>I want to say "16 bit" but I used them so long ago I am not certain. Last time
>I touched one was 1985 before I left USM and stopped teaching hardware design
>courses... as a result, I don't have any of the motorola poop sheets handy...
Not that this constitutes any real evidence, but back when Mephisto was making
the Dallas and Roma on 68000s they described them as 16-bit and the 68020 as
32-bit.
Albert Silver
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.