[Gtkradiant] ATI FireGL 8800 issues

[Timothee Besset]

On Tue, 15 Oct 2002 13:45:14 +0200
David Olofson <david.olofson@reologica.se> wrote:

> On Tuesday 15 October 2002 12:07, Timothee Besset wrote:
> > > 	* Xi Graphics have no driver for the FireGL 8800.
> > > 	  (ATI doesn't seem to care to give them specs and
> > > 	  permission... :-/ )
> >
> > Did you try DRI drivers? You might have better luck with them.
> 
> Well, AFAIK, they shouldn't work... At least, no other 8500 drivers will. 
> (Win98 drivers just crash, and XiG Summit won't start.)
> 

http://dri.sourceforge.net/dri_status.phtml
lists the 8500 in the supported chipsets

The rendering changes are new to 1.2, and have been a great improvement
overall (that is, on supported drivers). Spog wrote that code, he can
probably look into the fixes. It sounds like this is yet another fix we
will want in 1.2.11

TTimo

> 
> > > I grabbed the CVS tree last friday, and got it working by always
> > > setting both faces to the same mode. (Most importantly, GL_FILL for
> > > brushes, as that's what caused brushes to vanish in the camera view
> > > on the 8800.) I suspect that this "fix" also makes G400 cards run
> > > with h/w acceleration again.
> >
> > Do you mean you have to render the brushes in a different way to deal
> > with ATI driver bugs?
> 
> Well, just setting both faces to GL_FILL works. GL_FILL for one face and 
> GL_LINE for the other does not. (On the FireGL 8x00, nothing is visible - 
> it just eats fillrate. The G400 switches to dog slow s/w rendering.)
> 
> 
> > Is it valid for both platforms (win32 and linux)?
> 
> Yes, AFAIK. (Same problem with Win2k, although I haven't verified that my 
> "fix" works on Win2k. I also haven't verified that it really fixes the 
> "G400 gets dog slow" issue, but I'm quite sure it does.)
> 
> 
> BTW, when did GtkRadiant start using GL_LINE for backfaces? Must have 
> been some time after 1.1-TA, right? (That worked on G400/Win32, but no 
> 1.2.x verson does.)
> 
> 
> > A patch would be of interest.
> 
> Ok. I'll try to come up with something cleaner, though. Should be 
> possible to wrap qglPolygonMode() or something, I guess... (I've just 
> changed the camera code for now, which probably means that some other 
> stuff still fails to work on the 8800.)
> 
> 
> > Also, does turning on/off the use of display lists change anything.
> 
> Nope, at least not before the "fix". (Not tried with/without after 
> that.)
> 
> 
> > > BTW, I had to make a few minor changes to get GtkRadiant to build on
> > > my system. It might have to do with that system (still...) running
> > > Mandrake, but if you're interested, I'll post a patch.
> >
> > We are aware of some issues which we haven't fixed yet. ETG could give
> > you some info. We need to do some config checks prior to build I think.
> > Submit a patch anyway, it may be a good start point to settle this
> > cleanly.
> 
> Ok. (I have the stuff at home, though, so I can't post it right away.)
> 
> 
> //David Olofson - Programmer, Composer, Open Source Advocate
> 
> .- Coming soon from VaporWare Inc...------------------------.
> | The Return of Audiality! Real, working software. Really!  |
> | Real time and off-line synthesis, scripting, MIDI, LGPL...|
> `-----------------------------------> (Public Release RSN) -'
> .- M A I A -------------------------------------------------.
> |    The Multimedia Application Integration Architecture    |
> `----------------------------> http://www.linuxdj.com/maia -'
>    --- http://olofson.net --- http://www.reologica.se ---
> 
> _______________________________________________
> Gtkradiant mailing list
> Gtkradiant@zerowing.idsoftware.com
> http://zerowing.idsoftware.com/mailman/listinfo/gtkradiant
>