Re: PC Transporter OS

[mitch2gs@hotmail.com (Mitchell Spector)]

 "william strutts" <wrstrutts1@home.com> wrote:

>"Mitchell Spector" <mitch2gs@hotmail.com> wrote in message
>> pausch@saaf.se (Paul Schlyter) wrote:
>>
><snip>
>>     Windows 95 requires a full 32-bit CPU. The 386SX is not.
>> A 386DX CPU is the minimum requirement.
>>
>>>It can run 32-bit code just as well as the 386DX - the only difference is
>>>that the speed of execution will be slower on the SX.
>>
>>     Not so. The 386SX internal data bus is _16-bits_ wide and
>> its external data bus is 32-bits wide. This makes it more similar
>> to the 286; in fact it even shares the same 24-bit address
>> bus (16 MB total address space).
>
>I believe that the FPU was missing in the 386SX too.  The 386DX
>had the FPU.

    Actually no, there's no such thing as a 386 CPU with built-in FPU
(SX or DX). In fact I'm fairly sure even the AMD386-40 lacked an
FPU--I have a board with an AMD CPU surfaced mounted in place
and there's still a separate math co-processor socket.
   
>was the same processor.  The 386 SX and 386 DX were totally
>different processors.  Intel of course has repeated this with the
>Pentium versus Celeron's.  I have a Celeron III in my computer
>which is a crippled Pentium III with Coppermine core.  They
>reduced the cache in half and dropped the bus speed to 66mhz.

    Interestingly enough, the Celeron 300a outperformed the much
more expensive PII-450's when you ran it at 450/100 MHz FSB. :)
The Celeron II's and PIII's were a different story.

    Heh, how about getting this back on topic. Did you know the
65C816 was internally 8-bits wide while externally 16-bits wide?
That makes it a hybrid 8/16 CPU, much like the Intel 8088. I've
always wondered if for that reason it's technically correct to call
the Apple IIGS (or SuperNintendo) a true 16-bit machine. By
contrast the Amiga and first Macs used hybrid 16/32 68000's
and were considered 16-bit machines.

Mitchell Spector
a_specto@alcor.concordia.ca