Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!cs.utexas.edu!usc!jarthur!polyslo!vlsi3b15!batman!nicholaA
From: nicholaA@batman.moravian.EDU (Andy Nicholas)
Newsgroups: comp.sys.apple
Subject: Re: GS Wish List/Apple twoee compatiblity
Summary: bullshit
Message-ID: <958@batman.moravian.EDU>
Date: 3 Feb 90 20:01:04 GMT
References: <15474.chatter.infoapple@pro-beagle>
Organization: Moravian College, Bethlehem, PA
Lines: 68

In article <15474.chatter.infoapple@pro-beagle>, mmunz@pro-beagle.cts.com (Mark Munz) writes:

> What truly hurt the Apple IIGS computer is:
> 
> 1) the 65816 (at 2.8 Mhz) can't handle all the complex graphics.
>    Heck, the 68000 at 7Mhz can't!! A 68030 Mac SE/30 seems to
>    be a good minimum CPU system for handling such graphicly
>    intense environments.

That is just so much poppy-cock.

A 2.8 Mhz can handle the GS's current graphic capability just fine; the problem
lies within the environment which does all the drawing: QuickDraw II.  When
apple first wrote Quickdraw II, they didn't do the best job they could have
done (or we would have started with System 5.0, wouldn't we?).  If they
>had< done their very best, though, we probably wouldn't have had an operating
system for 3 years...

Also, it matters that many people write code which is just so much drivel and
usually very inefficient when doing graphics routines.  Many people/companies
like to write stuff with high level languages which are portable, but
on the IIGS this ends up meaning that your code will be slow.  If you invest
the time need to learn assembler on the IIGS, then the code which you write
is inherently worth less because it's not easily portable to anything other
than a 65xxx series machine.

> I would like to applaud the System 5.0 authors who have have
> pushed some of the 65816 limits -- and I usually only complain!!

System 5.0 doesn't push the 65816 to its limit.  hardly.  Look at any game
which scrolls a large section of screen around (like Alien Mind or Rastan)
and you'll see something which pushes the envelope.  The only program I
know of off the top of my head which -REALLY- pushes the machine right to
the edge is the opening screen to John's _Tomahawk/GS_.  I remember him
telling me that the opening screen is doing delta decompression (john's
own ace before ace was ace) while he is cycling all the pallettes to
show a 3200 color picture and updating the DOC.  He said he calulated that
he was using 97% of the total system cpu time.  I'm not sure if he was
using scan-loine interrupts or not in that one or if he was polling the
mega II video registers to check when to switch pallettes.

The spiel we got at this past may's applefest from the guys at apple was that
they just special-cased the routines which did the most-used object drawing
on the IIGS to speedup Quickdraw II.  Things like _PaintRect were made much
faster while routines like _PPToPort were left for future system disks to
improve.  Try moving a window around sometime.  The background screen
redraw is the same speed as system 4.0... so there is much more work that
could be done to speedup quickdraw II on a standard 2.8 Mhz IIGS.

Another for-instance comes from Jim Mensch at this past Kansas City 'Fest:
Jim was standing around talking about Quickdraw II and how to increase its
speedup and he said that the routines (I can't quite remember) for drawing
rects were something like 10X faster than system 4.0, but that he had just
written some stuff to get them to go 14X faster.  So, yes, there is more
than can be done to speedup Quickdraw II.

Part of the problem that I forsee apple facing if they try to radically
speedup Quickdraw II is memory: if they special case a lot of quickdraw II's
routines the special-casing is going to eat up a lot of memory.  And, who
wants to buy a 2 meg machine just to run one application?  Sure, it might be
as fast as a Mac II graphics-wise, but it's hard to justify putting 2
meg in a machine for that purpose increasing the cost to the end-user.

andy

-- 

Yeah!