Re: The joy of octal

On 11/16/24 11:16 AM, Don_from_AZ wrote:

"186282@ud0s4.net" <186283@ud0s4.net> writes:
 

On 11/16/24 12:24 AM, rbowman wrote:

On Fri, 15 Nov 2024 23:31:26 -0500, 186282@ud0s4.net wrote:
>

     Again, not entirely sure where the end of octal was. Many of the PDPs
     used octal, and I *think* a few PIC chips. 8/16/32 kinda took over
     kinda early on however.

chmod 4755
I don't know if I'd call it octal but if you were writing an
assembler for
quite a few microcontrollers the opcodes would have a pattern where source
ans destination registers were 0 - 7,

>
>
   Octal does persist, sometimes in obscure ways and places.
   It WAS kinda big for awhile - a "big step" better than
   8-bit.
>
   Alas don't think anymore 12 or 24 bit CPUs are
   gonna be made. Might still have a place for some
   higher-end microcontrollers - hell, I think Epson
   still makes FOUR-bit microcontrollers (looked at
   the sheet for one once, insanely capable).
>
   Hmmm ... 256 of those 4-bitters running
   parallel - that'd be a fun project :-)
>

 GE's "GECOS" and later Honeywell's "GCOS" mainframe machines were all
36-bit words, so octal was a natural for them: 6 6-bit BCD characters or
4 9-bit bytes per 36 bit word.

   Yep ... 2^12 hung on for quite awhile.
   And, in Linux/Unix, is STILL there in things like 'chmod'.
   8/16/32/64 seems more 'natural' ... but that may
   be more because of constant exposure than because
   of practical function. You can make a CPU with
   any word length you want.
   Remember "bit-slice" CPUs ? Fabrication tech could
   not make really wide single chips, so you just
   wired a bunch of them parallel ... you could HAVE
   yer 64-bit+ CPU even in the late 70s.