delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/1997/10/16/23:02:46

From: "Alexey Kouzmitch" <arexey AT gis DOT net>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: 3D hardware acceleration in DJGPP (Was: Re: allegro & 3d?)
Date: Mon, 13 Oct 1997 04:45:44 -0400
Organization: All USENET -- http://www.Supernews.com
Lines: 74
Message-ID: <61sn9u$hhd$3@e4000.supernews.com>
References: <Pine DOT GSO DOT 3 DOT 95 DOT 971008073505 DOT 9419A-100000 AT catbert DOT ucdavis DOT edu> <61llbt$jr3$1 AT news2 DOT xs4all DOT nl> <343e8ebd DOT 21393626 AT news DOT btinternet DOT com>
NNTP-Posting-Host: 17965 AT 208 DOT 218 DOT 132 DOT 248
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

 Allegro's 3d part IS slow!!!!
I took a look at the source for the polygon filler and it's COMPLETELY C...
for a good software renderer you must have the inner loop in
assembler!!!!...

I am using the original Allegro for now, but i am plannung on changing it
later to assembly.... i might even share it later :)

Alexey Kouzmitch..


Thomas Harte wrote in article <343e8ebd DOT 21393626 AT news DOT btinternet DOT com>...

>>Allegro does not make use of any of the hardware accelerated features
offered
>>by any 3D card.
>
> Let's just say hardware developers aren't particularly
>interested in helping what is commercially speaking a minority
>community. And then, if you are someone like S3, you aren't very
>interested in helping anyone at all :) At least that's what people have
>told me!
>
>>The 3D part of Allegro is not likely to be extended in the near future,
but
>>even for un-accelerated 3D, it's still a bit slow.
>
> It's not that it is so slow, it's that it doesn't do what it
>could. That is, it perspective corrects every pixel, I have heard people
>say it only clips on a pixel level, though I doubt that, and as for
>gouroud shading, I really don't know what's going on there, but it seems
>to be accurate to more than a fixed point scheme would allow. Maybe it
>is a brensham style maximum error thing, but I feel a fixed point scheme
>might be faster. Not that I'm any kind of authority!
>
>>You can try using another library such as Lib3D, Plush, Jaw3D, etc, and
get it
>>to draw the image to an Allegro bitmap (you would also have to tell the
other
>>library the height, width, and colourdepth of the bitmap, as it won't be
>>familiar with Allegro's BITMAP struct). AFAIK, none of these support
hardware
>>accelerated 3D.
>
> Just to say here, I am busy making a 3d library specifically for
>Allegro - no 3d hardware support yet since I don't own any (except an S3
>Virge, and I have a toolkit for that downloading now), but if anyone
>wants to get together on this, just de-spam my e-mail address (the '_'s)
>and say. It's not going badly, but I'll post an ANNOUNCEment when it
>gets anywhere decent. Go to http://www.softysoft.home.ml.org and
>download the gouroud man from the quickies section who I did in a half
>hour using a 0.1a version of my code if you want.
>
>>Are there any libraries out there for Dos/DJGPP that make use of hardware
3D
>>acceleration, or are there any under development? When I get a 3D card of
my
>>own, I might experiment with writing a DJGPP library for it, and perhaps
make
>>it possible to easily add support for other cards as well as
software-rendering
>>(for when there's no 3D card).
>
> Try MESA, it is OpenGL to all extents and purposes and so uses
>3DFX if it can. I don't know any URL's or anything. It works with DJGPP,
>but it is very unspecific and hard to figure out in my opinion. Anyway,
>if you ever do get 3d accelerator code going, any chance of a look? :)
>
> -Thomas
>.
>



- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019