home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!know!hri.com!noc.near.net!nic.umass.edu!caen!batcomputer!munnari.oz.au!bunyip.cc.uq.oz.au!news.qut.edu.au!qut.edu.au!podesta
- From: podesta@qut.edu.au
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: New hardware reference guide?
- Message-ID: <1992Nov13.085247.58456@qut.edu.au>
- Date: 13 Nov 92 13:52:47 GMT
- References: <Bx7782.CKD@kingston.ac.uk> <1992Nov5.184916.17436@sth.frontec.se> <1927@lysator.liu.se> <BxAMtH.Aos@cck.coventry.ac.uk> <36816@cbmvax.commodore.com>
- Organization: Queensland University of Technology
- Lines: 54
-
- In article <36816@cbmvax.commodore.com>, peter@cbmvax.commodore.com (Peter Cherna) writes:
- >
- > Well, we have extensive experience in figuring out why so many of these things
- > stopped working. We looked at a bunch that broke when a 1MB Agnus was
- > installed. We looked at a bunch that broke under ECS. We looked at a
- > bunch that broke with the A3000. A bunch that broke with 2.04. A bunch
- > more that broke with AGA and 3.0, and sadly, we fully expect to have the
- > same darned problem next time we rev something too.
-
- So what are some reasons that they break???
- I know some of mine broke because I wasn't blitter waiting everywhere.
- On faster machines, the processor would reach the next blitter job before
- the blitter was finished that last, while on A500s it would not and there
- would be no problem.
-
- >
- > From our experience, we've indeed seen some software that circumvents
- > the OS because that was the only way to achieve the effect desired.
-
- Definitely true in demos. If the OS could do it, there would be no
- point in making a demo out of it.
-
- > But by and large, the most common reason appears to be "NIH" syndrome (not
- > invented here). It appears to be common to underestimate the performance
- > of the system routines, to misunderstand how to use the system routines,
- > to misunderstand the intricacies of the hardware, etc.
- >
-
- Underestimate the performance of the system routines?????????
- Come on, is that possible. True, that 2.x is much better than 1.x and
- 3.x sounds great, but there's still NO way your gonna write a game
- like Shadow of the Beast III that will run on a 1meg A500 using system routines!
- And to compete with the Console games, and to get people to buy your games,
- and just to feel proud of your games, you need to target that kind of level
- of speed, graphics and gameplay. There's no market for games and demos that
- only run on 68020 machines and above. Even when there is, this just
- means people will expect better demos and games so we still won't be able
- to go through the OS
-
- > We frequently see code that goes to the hardware that is significantly
- > less efficient than it could easily be, or even than the OS is. The
- > only thing which is rare in all of this is code which takes over the machine
- > completely and DOES work on all Amigas. Perhaps you are among those who can
- > write code that works on most Amigas, but you're in the tiny minority...
-
- In demos??? They must be pretty lame if some hardcoded effects are slower than
- the system software...Remember that demos and games can use the fact that
- they are not generic - we don't have to worry about window clipping or
- multiple tasks or I/O because these are all controlled.
-
- >
- > Peter
-
- James Podesta. ((soon) BappSci - Computing Honours :) a.k.a Rincewind
-