Path: utzoo!attcan!uunet!lll-winken!sun-barr!cs.utexas.edu!sdd.hp.com!samsung!zaphod.mps.ohio-state.edu!swrinde!ucsd!ucbvax!hplabs!hpcc05!hp-ptp!marcb
From: marcb@hp-ptp.HP.COM (Marc Brandis)
Newsgroups: comp.sys.intel
Subject: Re: Difference between a 386 and a 386sx
Message-ID: <1310001@hp-ptp.HP.COM>
Date: 26 Sep 90 16:28:36 GMT
References: <1990Sep16.194605.11968@ecn.purdue.edu>
Organization: HP Pacific Technology Park - Sunnyvale, Ca.
Lines: 19

Yes, there is a way to distinguish, but it involves measuring the
time. Just do a long DWORD string instruction (like STOSD), do it
once aligned on a dword boundary and once aligned on a word boundary
but unaligned on a dword boundary (that is at the same address+2).
Time both runs. On a 386DX, the time for these two access sequences
differs quite a lot, as it needs one access for each doubleword in
the aligned case and two in the unaligned case. On the 386SX, the
times are the same as it requires two bus cycles anyway.


(* I speak only for myself.
	Marc-Michael Brandis
	Institut fuer Computersysteme
	ETH Zentrum
	CH-8092 Zuerich, Switzerland
	e-mail: brandis@inf.ethz.ch
		brandis@iis.ethz.ch
   Temporarily at HP, marcb@hp-ptp.ptp.hp.com
*)