Path: utzoo!mnetor!uunet!mcvax!ukc!stl!stc!datlog!slxsys!jpp
From: jpp@slxsys.specialix.co.uk (John Pettitt)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Norton si - how  does it determine the cpu?
Message-ID: <110@slxsys.specialix.co.uk>
Date: 16 Dec 87 09:52:29 GMT
References: <707@cblpe.ATT.COM>
Reply-To: jpp@slxsys.UUCP (John Pettitt)
Organization: Specialix International, London, UK.
Lines: 34
Summary: Count the prefetch Q

In article <707@cblpe.ATT.COM> res@cblpe.ATT.COM (55216-Rob Stampfli) writes:
>I have an AT&T PC6300 with a NEC V30 processor.  When I run the Norton Utility
>program "si", it sometimes reports the processor chip as a V30, and sometimes
>(maybe 25% of the time) as a V20.  Does anyone know how Norton decides what
>chip is being used?  Has anyone else noticed this problem?

I Don't know how SI does it but here is a good way to find out
what CPU you have.

1) execute at least 30 identical short instructions - this
ensures the prefetch Q is full.

2) Use self modifying code to change an instuction n bytes
ahead to an inc or a flag clear or the like.   

3) If the changed instruction is executed it was not in the Q
decrement n and try again.

4) if the changed istruction was not executed that location had
already been prefetched.  Therefore the Q is n instructions.

The Q on 80[88,86,188,186,286,386] is unique length for each
chip.  (Sorry I don't have the lengths).  I goes without
saying that interrupts must be disabled for correct operation.

This method was, I believe, invented by Jude and Howie at 
Applied Reasoning Corp, Cambridge, Mass.  



-- 
John Pettitt - 144.5 MHz: G6KCQ, CIX: jpettitt,  Voice: +44 1 398 9422
UUCP:  ...uunet!mcvax!ukc!pyrltd!slxsys!jpp  (jpp@slxsys.specialix.co.uk)
Disclaimer: I don't even own a cat to share my views !