Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!think!snorkelwacker!bloom-beacon!eru!luth!sunic!tut!santra!mcsun!ukc!acorn!moncam!emmo
From: emmo@moncam.co.uk (Dave Emmerson)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Xerox sues Apple!!!
Summary: Not these days..
Message-ID: <336@marvin.moncam.co.uk>
Date: 29 Dec 89 16:53:47 GMT
References: <6767@tank.uchicago.edu> <1989Dec17.112127.27333@me.toronto.edu> <332@sixhub.UUCP>
Organization: Monotype ADG, Cambridge, UK
Lines: 45

In article <332@sixhub.UUCP>, davidsen@sixhub.UUCP (Wm E. Davidsen Jr) writes:
> [much deleted]
>   The hidden advantage of the Intel instruction set is that many of the
> instructions are single byte, thus reducing memory size (minor benefit)
> and allowing a greater number of instructions to be fetched per unit
> time given a limited memory bandwidth. This makes a difference in
> performance at the low end, with slow memory.
> 

Well that used to be true when 8 bit bus machines ruled, but these days
16 and 32 bit busses are rapidly replacing them. Bear in mind that the
68008 was actually a cut-down 16/32 bit processor, not really an 8 bit
predecessor of the 68000, so 16 bit opcodes are more 'natural' for this
range. 8 bit opcodes would actually be wasteful on the 68000 and 68010,
and the 68020 had to implement some *really neat* circuitry to allow
random 8 bit data fetches on a 32 bit data bus without slowing the
processor to shift it into place. 8 bit opcodes would have absolutely
*NO* advantage whatsoever unless you shackled the '020 to an 8 bit bus,
and this is usually only done for 'bootstrapping'.
Perhaps someone in the know can tell us how the '486 compares with the
'020s dynamic bus sizing?

>   I am NOT proposing any Intel or other CPU as "better" in some way
> than another, just adding a few observations. 

Ditto. I cut my teeth on the SC/MP, hacked the 6502 and Z80, and *loved*
the 68k, but nowadays I need the portability of C. I'd probably have
hacked '86 assembler if it were not for those disgusting segments. Yuk!
OK, yes, I admit it, I do have a preference, and it is heavily biased,
but '86 looks too much like a step backwards. It's bad enough having to
use 'far' in C to make some progs run on a PC. I would hope that this
botch will soon disappear from all PC C compilers, then perhaps we can
pretend that our machines are perfect, no matter what the processor...


Just MHO, the world would be a dull place if we all agreed.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

static char disclaimer[] = { 
"These are MY opinions. My employers don't have the collective IQ to form
 an opinion of their own, that's what they hired me for!" }

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~