View Single Post
  #7   Report Post  
Posted to rec.crafts.metalworking
DoN. Nichols DoN. Nichols is offline
external usenet poster
 
Posts: 2,600
Default printer pieces anyone

On 2010-07-19, wrote:
On Jul 18, 7:22*am, "Bill Noble" wrote:

sometimes you need hard copies - like for friends with no computer, or to
read and mark up - but yes, printers have certainly lost value - this one
was over $800 new. *I'd sure like to find a home for some of the parts - the
main board has a 68000 processor on it - remember those?


I sure do. What a slow POS. Wanted to be 32 bit, but had only an 8 bit
bus.


Huh? Try a 16-bit bus. I think that the 68008 had the 8-bit
bus, but the 68000 was a full 16-bit bus.

How about comparing it to the Intel 8088 used in the original
IBM PC about that time? *That* one *was* stuck with an 8-bit bus.

I had a unix based system running on an 8 MHz 68000. While it
was slow -- the worst thing about it was the compiler. It produced code
as though it were for a PDP-11 instead of a 68000. The only
68000-unique codes which it bothered to use were "LINK" and "ULINK" (for
building and destroying stack frames). Otherwise, if the PDP-11 didn't
have it, the compiler didn't use it. :-) There were a lot of CISC
instructions in the 68000 (and later) which could have speeded up the
code significantly.

In particular -- I saw output from the compiler doing access to
a two-dimensional array by going through multiple steps to calculate
the offset from the base address of the array. Lots of calculations,
when you have to take into account the size of the array elements among
other things.

The later 68000 family included instructions which would do it
all in a single instruction -- base address in one register, size of
data elements in the instruction, size of a row in a register or in the
instruction, row and column offset in the instruction or in registers,
and bingo -- it could read or store in the proper element without all of
the extra instructions.

I loved the 68010 and 68030.


The 68010 had the same bus size as the 68000 -- 16 bits. The
68020 took the bus size to 32 bits, and the 68030 put the floating point
hardware math on the chip too.

The ones I programmed were in UNISYS
high speed check reader/sorters.


O.K. My 68000 system was a Cosmos CMS-16/UNX -- built in an
Intel Unibus card cage. 8 MHz CPU Clock. It included the memory
management chips right beside the CPU. The OS port was by Unisoft.

Later systems which I had which used the 68000 family we

1) 68010 -- AT&T Unix-PC (AKA 3B1) -- a cute desktop machine
with a futureistic style. 10 MHz CPU SysV flavored unix.
Rather limited as supplied, but lots of people hacking them to
improve them, such as allowing a second disk drive, and allowing
both to exceed 67 MB total size (ST-506/MFM style drives). The
largest drives which could be connected (based on availability in
that interface style) were 190 MB drives by Maxtor and one other
brand.

This machine, and the following one, were both significantly
faster than the 68000 machine -- which I blame on the poor
compiler used on the 68000 system, not the CPU's instruction
set.

2) 68010 -- 10 MHz CPU Sun 2/140 (Multibus cage, BSD flavored unix.

3) 68020 -- Several members of the Sun3 family BSD flavored unix

Now -- a Sun system which I did *not* ever have:

68030 -- Sun3x family -- one desktop machine in the same
physical format as the Sparc1, and one pedestal server machine.

By the way. All the printers I have ever junked out had one or more
very nice ground and polished stainless steel rod. Very usable to make
metal things!


Indeed so. Ground and polished -- and usually hardened, too.

Enjoy,
DoN.

--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. |
http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---