Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!shadooby!accuvax.nwu.edu!tank!uxc!deimos.cis.ksu.edu!rutgers!apple!oliveb!amiga!cbmvax!thomas
From: thomas@cbmvax.UUCP (Dave Thomas QA)
Newsgroups: comp.sys.amiga
Subject: Re: Software vs 1M CHIP RAM
Message-ID: <6719@cbmvax.UUCP>
Date: 1 May 89 13:30:48 GMT
References: <781@manta.NOSC.MIL> <1fsJ489nAw1010C8s.w@amdahl.uts.amdahl.com> <526@laic.UUCP>
Reply-To: thomas@cbmvax.UUCP (Dave Thomas QA)
Organization: Commodore Technology, West Chester, PA
Lines: 26

In article <526@laic.UUCP> darin@nova.UUCP (Darin Johnson) writes:
> Hmmn.  Just how does a program go about making itself NOT work with
> 1M chip memory?  The only thing I can think of is if the program
> checks to see if allocated memory is really chip memory by looking at
> the address; but of course no-one would break such a taboo :-)
> Any other methods?  Just curious.
> 

Well...a couple of ways.  The first involves making assumptions about
the configuration of the machine and when those assumptions don't hold
true the software gets confused.

The second involves direct register manipulation of the custom chips.  Many
bits were specified as reserved/unused on the old chips.  Programs were 
supposed to never write into these bits.  Even if they did, the old chips
ignored these bits and no harm was done.  The new chips use some of these
bits and programs which blithely write to them can cause some interesting
visual fireworks.

You are most likely to see problems with games and video programs which often
directly manipulate the registers for performance reasons.

        Dave
-- 
Dave Thomas, Commodore Amiga Test Engineering
UUCP  ...{allegra,rutgers}!cbmvax!thomas