[comp.windows.ms] no subject

dev@chinet.chi.il.US (Lawrence Weeks) (06/16/91)

Newsgroups: comp.windows.ms
Subject: Re: Mac Vs. Windows? (sorry)
Summary: 
Expires: 
References: <0E010021.e0mxxc@gla-aux.uucp> <1991Jun4.154854.19649@dbase.A-T.COM> <56nD02AQ08cd01@JUTS.ccc.amdahl.com>
Sender: 
Followup-To: 
Distribution: 
Organization: Chinet - Chicago Public Access UNIX
Keywords: 

In article <1991Jun4.154854.19649@dbase.A-T.COM> tomr@dbase.UUCP (Tom Rombouts) writes:
>In article <0E010021.e0mxxc@gla-aux.uucp> glenn%gla-aux.uucp@skinner.cs.uoregon.edu writes:
>>
>>In article <1991May18.050842.5732@cs.uoregon.edu>, akm@obelix.cs.uoregon.edu (Anant Kartik Mithal) writes:
>>> So, my question is: if the Mac is faster than a "comparable" pc in
>>> execution, what is the reason? Is it ROM? (I don't think so for the
>>> reasons listed above.) What other things can make a difference?
>>
>>Two things -- processor effeciency and levels of routines.
>
>[ rest of post deleted ]
>
>>2) The Motorola (and for that matter, any linear-addressing processor) is
>>faster than segmented memory.

One must keep in mind that Motorola architecture has always been superior
to Intel architecture. This is the widely accepted opinion in the design
circles, and is why introductory microprocessing is taught on something from
Motorola (e.g. 6809) rather than Intel in the good EE schools. I've heard
professors laugh about the Intel 80x8x family. :-) Granted, Intel may have
finally caught up with the 80386. But it retains the stilted opcodes, etc...
from the previous members of the family. A Mororola processor running at the
same clock speed as the "equivilent" Intel processor will be faster, because
the Motorola executes instructions more efficiently. Intel's solution is to
pump up the MHz on their CPUs to overcome the inefficiency.

If only IBM had contracted with Motorola's 6809 (still considered the best
8-bit chip by many) back then rather than Intel's 808x family, it would be
a vastly different and better PC world. Hell, my old 2 MHz 6809B handles
512K of RAM and multi-tasks in windows more solidly and smoother than
MSWindows 3.0 does. And maybe now PC users wouldn't be saddled with umpteen
versions of an OS (MS-DOS) that's so bad the software writers choose to
ignore it rather than work with it. And one so bad that it gives them that
option too. A good OS will CONTROL the machine, not let any old piece of
software come up and take over. I have a friend who knew one of the authors
of MS-DOS. He tends to make jokes about him.

Oh well, what is done is done. Only now are the PC users starting to climb
out of the dark ages induced by the mating of Intel and Microsoft's
incompetency. At least OS/2 2.0 (from IBM, not Microsoft) looks promising
as a real OS. Windows could have been nice if they had abandoned the DOS
foundation. Writing something for DOS is like building a house on top of
an active tectonic fault line. Such is the PC world.

Larry

mcdonald@aries.scs.uiuc.edu (Doug McDonald) (06/16/91)

In article <m0joeFD-0000v7C@chinet.chi.il.us> dev@chinet.chi.il.US (Lawrence Weeks) writes:
>
>
>One must keep in mind that Motorola architecture has always been superior
>to Intel architecture.


WOW!!! What a sweeping statement.  **WHY** is it superior??



>This is the widely accepted opinion in the design
>circles, and is why introductory microprocessing is taught on something from
>Motorola (e.g. 6809) rather than Intel in the good EE schools. I've heard
>professors laugh about the Intel 80x8x family. :-)

Now we know - the 68... is superior for teaching elementary courses.
Now that, perhaps, I'll agree. HOWEVER - does that make it superior
for building computers??? Not necessarily. There are many things
besides elegance that matter there. Like how well it fits in with the
rest of the hardware, how fast it is, how cost effective, etc.



>Granted, Intel may have
>finally caught up with the 80386. But it retains the stilted opcodes, etc...
>from the previous members of the family.

Why cares?? I seldom write machine code. Most people NEVER write machine code.
[Yes, I have and still do sometimes write machine code. For those
academic heads who don't know what "write machine code" means, it means
to look up the numerical values of op-codes and operand codings in the
processor hardware manual and code the bits yourself. Somebody has to do
this, you know.]

>A Mororola processor running at the
>same clock speed as the "equivilent" Intel processor will be faster, because
>the Motorola executes instructions more efficiently. Intel's solution is to
>pump up the MHz on their CPUs to overcome the inefficiency.
>

For the 680x0 and 80y86 series the is NOT SO if y == x. I used to think that 
the 68040 had caught up with the 80486, then along came two more benchmark
tests that showed the 80486 faster even for floating point.


YOU CANNOT MAKE VAILD CLAIMS THAT THE 680x0 IS FASTER THAN THE  80y86
WITHOUT BACKING THEM UP WITH REAL BENCHMARKS ON REAL PROGRAMS ON REAL
COMPUTERS. And this is impossible to do for x == y < 4 and not conclusive
for x == y == 4.


The IBM PC is the most successful computer the world has ever seen. 
An absolute majority of all computers are based on the Intel family.
It is now almost to the point where more Intel computers are SOLD every year
than the TOTAL of all other types sold in all previous years. This did NOT
happen by accident: it happened because the PC was the right design for its 
time - it was efficient and cost effective. There WERE, you know, Motorola
based computer for sale then. They are on the ash heap of history now.
Only one maker (Apple) has ever sold a significant number of Motorola
computers and their sales are minute compared to those of PCs.


IF the original poster wishes to convince me that the Motorola chips
are superior, he is welcome to point me at standard, REAL, benchmarks,
that show it is superior. I also point out that I accept comparisons
of two sorts: two computers that cost the same thing, OR two computers,
each of which is currently the absolute fastest using the given chip. 
It is required to use the best available compiler for each chip.

It is up to him to prove it. 


Doug McDonald

mtang@sgi.com (Man Kit Tang) (06/20/91)

In article <m0joeFD-0000v7C@chinet.chi.il.us> dev@chinet.chi.il.US (Lawrence Weeks) writes:
>One must keep in mind that Motorola architecture has always been superior
>to Intel architecture. This is the widely accepted opinion in the design
>circles, and is why introductory microprocessing is taught on something from
>Motorola (e.g. 6809) rather than Intel in the good EE schools.

I think a cleaner architecture is actually more easy to understand than
a messed-up one.

>I've heard
>professors laugh about the Intel 80x8x family. :-) Granted, Intel may have
>finally caught up with the 80386. But it retains the stilted opcodes, etc...
>from the previous members of the family. A Mororola processor running at the
>same clock speed as the "equivilent" Intel processor will be faster, because
>the Motorola executes instructions more efficiently. Intel's solution is to
>pump up the MHz on their CPUs to overcome the inefficiency.

My experience tell me otherwise, Intel processors starting from 386 (even
though still an architecture disaster) actually execute most integer
intensive programs much faster than equivalent Motorola processors (i.e.
386 vs 030 and 486 vs 040) at the same clock speed.  Motorola processors
however execute floating codes much faster.  These observations are under
UNIX of course so that the 386/486s are actually used as 32-BIT processors !

-mkt

P.S.  Standard disclaimer that the above opinions are solely mine and have
      nothing to do with my company!