Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!samsung!uakari.primate.wisc.edu!sdd.hp.com!decwrl!elroy.jpl.nasa.gov!zardoz.cpd.com!tmiuv0!rick
From: rick@tmiuv0.uucp
Newsgroups: comp.sys.amiga.hardware
Subject: Re: Motorola vs Intel (who's faster?)
Message-ID: <3161@tmiuv0.uucp>
Date: 12 Jul 90 11:21:23 GMT
References: <1990Jul10.055108.22796@agate.berkeley.edu>
Organization: Technology Marketing Inc., Irvine, CA
Lines: 60

In article <1990Jul10.055108.22796@agate.berkeley.edu>, joechung@sandstorm.Berkeley.EDU (Joseph) writes:
> A friend of mine said the other day that Motorola rates its CPU's twice as
> fast.  So in effect, my Amiga's running at a little over 3.5 MHZ.
> 
> One of the reason he sights is that: Motorola & Intel are about the same
> level in technology, and yet Motorola has come out with a 50MHZ chip already
> while Intel is still around 33MHZ.  Therefore, the actual speed of the
> Motorola chips are 1/2 what the manufacturer says.
> 
> No flames please.  I just want to know if what he said is true.  I'm a little
> surprised the hear that my 68000 actually runs at 3.5 compared to an Intel.
> 
> -jc
> =================================================================
> | Joseph Chung                  |  This area under construction |
> | joechung@ocf.berkeley.edu     |                               |
> =================================================================

Heheheh!  I take it that your friend is an Intelite.  No, that's not correct.
The clock speed for the Motorola parts are legit.  If you have a 50Mhz
68040, you feed it a 50Mhz clock.  If you have a 33Mhz 80386, you feed it
a 33Mhz clock.

He may be referring to throughput, which is the amount of work that a
processor can do in a fixed length of time.  Most commonly, this is
referred to in MIPS (Millions of Instructions Per Second) or using some other
benchmark such as Dhrystones (a measure of computational power).  Due to
architectural differences between processors and the way they're used in
a given computer, these measurements all have to be taken with a grain of
salt, particularly MIPS ratings.  The most common misuse of MIPS is comparing
a RISC processor (such as the Sun SPARC or Motorola 88000) against a CISC
processor (such as the 680x0 or 80x86).  True, most RISC processors have
a higher MIPS rating (they execute more instructions per second), but each
instruction on a RISC does far less work, requiring more instructions to
be executed to perform a particular task than a CISC processor would.

Better benchmarks are things like Dhrystones, which measure how many
computational results per second you get, and essentially disregards
RISC/CISC differences.  It's the results and when they get there that matter
to me, not now the results got there.

Oh, by the way, in case you don't know these abbreviations...

    RISC:  Reduced Instruction Set Computer
    CISC:  Complex Instruction Set Computer
 
For your own peace of mind, computationally, an Intel 80386 at 33Mhz has
roughly the same computational power as a Motorola 68030 at 25Mhz.  I have
not proven this absolutely, but in my experience, it's a fairly close call.

I MUCH prefer the Motorola method of doing things, from a programmer's point
of view (and that is what I be -- a programmer).
----------------------------------------------------------------------------
[- O] Rick Stevens
  ?   EMail: uunet!zardoz!tmiuv0!rick -or- uunet!zardoz!xyclone!sysop
  V   CIS: 75006,1355 (75006.1355@compuserve.com from Internet)

"I'm tellin' ya, Valiant!  Da whole ting stinks like yesterday's diapers!"
                                - Baby Herman in "Who Framed Roger Rabbit"
----------------------------------------------------------------------------