Warning re: unofficial drivers.

Literatii

New member
I was working on a Radeon8500 OpenGL project today and ran into a strange rendering problem with the 9016 and 9017 drivers. It was cleared up by reverting back to the 9009 drivers.

The creator of the component I was using summed up the problem after I tried various coding changes:

"Looks like some not-safe optimization is made by the driver when
compiling the triangle strip into a build list, and fails when that
strip is a degenerated one... Probably something introduced in the
later revisions or for the 8500 (just tried for ATI Rage Pro and 128
- no issue)."

Happy Coding.
 
Have you tried anything newer? 6037 is official beta afaik, and there are leaked 6043 or 6052 for you to try too.
 
...

...

I just tried the 9021 drivers, and the problem remains. I've also run into other various problems in the past few days with textures, and hard lock ups using the stencil buffer.

I've sent what I can to [email protected] but I'm fast deciding whether it's worth keeping this card. I don't want to spend the bulk of my development time coding around buggy drivers.
 
Well, in case you want some help I think it would be a good idea to either try to explain the problem in a useful way. <edit>

Edit: Ahem, I were wrong at some points, 90xx is the Win9x drivers, 9021 is the latest ...
 
Last edited:
...

...

I've sent the appropriate information to ATI regarding the buildlist compile error for degenerated strips. I don't think there's much room for discussion there: it just doesn't work on the 8500 with their drivers.

If you have any clue as to the other problem (GLSELECT) I'd be happy to hear it. I have another couple as well, but haven't pinpointed the exact causes yet.

--Lit
 
Back
Top