khaqq wrote: > On Wed, 16 Apr 2008 04:34:17 -0700 (PDT) > smoki <[EMAIL PROTECTED]> wrote: > >> This is happen in r200 driver, because when i use software TCL pipeline it's >> correct. >> Is there anybody who can handle this? > Ok, I've verified this. This is standard z-fighting due to a tcl fallback (so in the fallback path the fragments generated do not match those generated without fallback). You can see this with R200_DEBUG=fall, there will be messages about texgen fallback. In fact the fallback (in this case) seems unnecessary (tried it), but it's hard to tell since there are situations where it might be necessary. The code does its best to figure out what to do but there's some guesswork involved there. A fix would be possible, if nothing else it would be possible to convert the fixed function tnl into a vertex prog and run that so we wouldn't need a fallback (though I wouldn't know if this one would fit into the r200 resource limits), but this currently can't be done.
It shouldn't cause instabilities however, if it does maybe we're not handling switching on/off tcl correctly (maybe forget a vap flush or something), but I didn't experience that. Roland ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
