Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!akgua!gatech!seismo!brl-tgr!tgr!cottrell@nbs-vms.arpa
From: cottrell@nbs-vms.arpa (COTTRELL, JAMES)
Newsgroups: net.lang.c
Subject: 32 or 16 bits?
Message-ID: <1555@brl-tgr.ARPA>
Date: Mon, 13-Jan-86 15:10:03 EST
Article-I.D.: brl-tgr.1555
Posted: Mon Jan 13 15:10:03 1986
Date-Received: Wed, 15-Jan-86 00:47:08 EST
Sender: news@brl-tgr.ARPA
Lines: 47

/*
>     It seems to me that the arguments around 16-bit vs. 32-bit 
> architectures have gotten a little out-of-hand.  

Most assuredly.

> There are two DIFFERENT standards by which to judge processors.  

At least.

> First, the width
> of the physical data bus.  By this standard, the 8088 is an 8-bit
> processor, 8086 and 68000 16-bits, and the 68020 32-bits.

Many people (including me) disdain this standard. However, it is
easy to measure and the results (by this standard) are nondebatable.

>     The second standard is the internal organization of the processor.
>My own definition for this standard is the ability of the processor to
>execute ALL of its data move, logical, and arithmetic instructions on
>a given word size. ( Note.  I emphasize that this is MY standard. )
>By this criterion, the Intel products up to and including the 80286 are ALL
>16-bit processors, as are the 68000 and 68010. (They come REAL close to 32,
>but fall down on the MUL instructions).  The 68020 is a real 32-bitter in both
>cases.

I think you are being a little hard. Consider that the 8-bit micros have
NO multiply or divide at all (except the 6809 which has 8 bit multiply 
only). By your definition they are zero bit machines. Or maybe if it had
NO multiply at all it would qualify as a 32 bit machine. Consider further
that multiply & divide are the only primitive operations that return
a different number of bits than they consume (anticipating the counter-
example of `convert short to long' etc., I dismiss it with a wave of the 
hand). Consider lastly what percentage of instruxions are axually
multiply & divide. I think the 680[01]0 come close enuf to be called 32 bit.
And the 6800 is a 16 bit. But then, that's just *my* opinion.

>     The point of all this is that there are several ways of looking at the
>issue, any given one of which is correct, depending on what you're looking
>for.  I don't mind hearing discussions, even religious ones, but die-hard
>dogmatism is never comfortable for those on the outside.

What is religion besides diehard dogmatism?

	jim		cottrell@nbs
*/
------