home *** CD-ROM | disk | FTP | other *** search
- Path: Hermes.grace.irl.cri.nz!maths!peterm
- From: peterm@maths.grace.cri.nz (Peter McGavin)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Demo/game to OS frien
- Date: 05 Feb 1996 01:31:39 GMT
- Organization: Industrial Research Ltd
- Distribution: inet
- Message-ID: <PETERM.96Feb5143139@tui.maths.irl.cri.nz>
- References: <4e8h9j$mp5@sinsen.sn.no>
- <4e8pk2$ntm@serpens.rhein.de><3873.6603T379T952@wr.com.au>
- <PETERM.96Jan31172831@tui.maths.irl.cri.nz>
- <3274.6607T382T680@wr.com.au>
- NNTP-Posting-Host: tui.grace.cri.nz
- In-reply-to: accolyte@wr.com.au's message of 3 Feb 1996 10:04:51 GMT
-
- accolyte@wr.com.au (Accolyte) writes:
- >> Do you know how to handle interrupts from my GVP Spectrum card? From
- >> my WarpSCSI? From my network card? From my brother's OpalVision or
- >> Z3FastLane? From MPEG cards? From hardware that hasn't been designed
- >> yet (like new Amigas)? Disabling any of those interrupts for too long
- >> can cause all sorts of nasty things to happen, like lockups, network
- >> failures, corrupt displays and/or corrupt hard disks.
- >
- >Ah so that's what he was getting at. Like it or not this is rare though.
-
- Well it happens to people all the time. And the hardware is not to
- blame.
-
- >For compatibility one might consider using the system interrupt routines
- >as an *option* to the user, but taking over the interrupts should always
- >be available for those who want speed.
-
- Unless we're getting thousands of interrupts per second (which
- indicates we're probably using the wrong algorithm), no-one will
- notice the difference. Oops, yes they will --- it makes the
- difference between work/crash on some configs.
-
- >I'm just the opposite, I tend to forget about system ones. ;) If a
- >hardware bashing game doesn't run, it's the coder's fault not the fault
- >of the concept of hardware bashing in general. I've already mentioned
- >a few of the above in another post, but:
-
- I agree it's the coder's fault. The OS allows hardware bashing after
- checking-for and allocating the hardware. Of course this is at the
- expense of compatibility with other hardware.
-
- >Diamond Caves: Too slow, and plays nowhere near as well as BaldersGrove
-
- Could it be the high-res interlace mode that slows Diamond Caves down?
- In any case, direct bitmap access is possible without killing the OS
- and Diamond Caves probably doesn't have an option for that. I wonder
- whether BaldersGrove works at all on my system.
-
- >Gloom Deluxe: Point. Although it requires a much faster system to run
- > acceptably. The stock 1200 routines are not system ones.
-
- Gloom Deluxe would be faster on an A1200 if it didn't kill the OS and
- if it used QBlit() for c2p.
-
- >Spectrum Emul: Not a game :)
-
- So what if it's not a game. It's 10000 games, many of which are
- better than ones in this list (arguably).
-
- >Colonization: Shocking code, and this in terribly slow, even for OS calls.
-
- So how does killing the OS improve shocking code?
-
- >DuneII: Could be much faster
- >Angband: If it's what I'm thinking of (moria clone?) then that's
- > on par with the ol' text adventures in terms of system
- > requirements :)
-
- And people write text adventures and disk mags that kill the OS :-(
-
- >F18Interceptor: I can't remember this, but I'm sure it could've been much
- > faster/better.
-
- F18 Interceptor was written when A500s were the norm and it's still a
- classic, fast flight simulator. I can run my network server in the
- background and I can't tell when someone accesses my hard disk unless
- I look at the disk light.
-
- >Poing: I admit, this is a downright trendy/addictive game :)
- > But it's not overly complex is it?
-
- It would have to be incredibly complex before I would consider killing
- the OS. Far too many coders decide to kill the OS from the start.
-
- IMHO all the above games are better than you make out. On the other
- hand, none of them are perfect. Some use only high-level OS calls,
- hence they work with gfx-cards but are relatively slow on stock
- A1200s. This problem can be worked around without killing the OS.
- Indeed, all the games above could almost certainly all be improved
- without killing the OS.
-
- >I think the trend is that the ONLY OS programmed games that will run
- >acceptably are the simple ones, and the ones designed for graphics
- >cards only. Where does that leave the people who don't have/want a
- >graphics card?
-
- To work with all gfx cards, a game must use high-level OS calls. If
- that's too slow on a particular chipset then the game can allocate and
- bang within OS rules. A good game would give the user both options.
- --
- Peter McGavin. (p.mcgavin@irl.cri.nz)
-