Path: utzoo!attcan!uunet!fernwood!portal!apple!uokmax!d.cs.okstate.edu!ong
From: ong@d.cs.okstate.edu (ONG ENG TENG)
Newsgroups: comp.sys.ibm.pc.misc
Subject: Norton S.I. 3.1 is cheating!
Message-ID: <1991Feb1.054859.7614@d.cs.okstate.edu>
Date: 1 Feb 91 05:48:59 GMT
Organization: Oklahoma State University
Lines: 27

Well, well.  In playing around with the optional oscillator slot on
my 386SX motherboard, I found something very interesting with
Norton S.I. 3.1 program.

To make sure that the oscillators (32MHz, 35MHz, and 40MHz for
16MHz, 17.5MHz, and 20MHz operations respectively) are all working
properly, I wrote a SPEED program in assembler.  This is the
readings I am getting:

Operation     Oscillator       Norton S.I. 3.1           SPEED
=========     ==========       ===============           =====
  8.0MHz                            7.1                    7.1 
 16.0MHz         32MHz             15.3                   14.6
 17.5MHz         35MHz             15.3                   15.7

Notice that at 16MHz and 17.5MHz, both Norton S.I. 3.1 returned
15.3 whereas my SPEED program returned the correct proportional
increase from 14.6 to 15.7.  I suspect that Norton has a 
table built-in.  Parameters of the table include type of
machine (e.g. 286, 386DX, 386SX, etc), standard 
cpu speed (e.g. 8MHz, 10MHz, 12MHz, 16MHz, 20MHz, etc).
The program found that 17.5MHz is closer to 16MHz than 
20MHz, so it returned the 16MHz number.  Anybody has
any idea?  (my SPEED program uses LOOP command to delay against
the clock retrieve from INT 1ah).

E. Teng Ong (ong@d.cs.okstate.edu)