[Mesa-dev] [PATCH] lp_build_lod_selector: Disable brilinear folding on explicit lod.

[Jose Fonseca]

----- Original Message -----
> On Mon, Jun 25, 2012 at 03:16:35PM -0700, Jose Fonseca wrote:
> > Indeed lp_build_brilinear_lod is not faster than
> > lp_build_ifloor_fract, but brilinear is faster, not because log is
> > a
> > faster approximation, but because it increases the odds that fract
> > part is zero, which means that we can sample from a single mip
> > level, instead of lerping between two mip levels.
> > 
> > I think you have a good point here -- lp_build_brilinear_lod is a
> > log2 approximation which is wrong here and that's a great catch, --
> > but I have a point too: lp_build_ifloor_fract will slow down
> > texture
> > sampling.
> >
> > Just like log2 and brilinear log2, we need a variant of
> > ifloor_fract, that increases the probability of fract part being
> > zero, essentially by applying a stair case transformation like:
> 
> You can do that by multiplying by 'k', subtracting 0.5*k and clamping
> to [0,1[. 

Actually, looking at the code, this is precisely what lp_build_brilinear_lod should be doing (albeit a bit more efficiently).  It does not do a log2, as you implied earlier. log + brilinear is done by lp_build_brilinear_rho and not lp_build_brilinear_lod.

Olivier, please look at the failure again, and confirm exactly what's the problem: does lp_build_brilinear_lod have a bug, does the assumption that clamping is not necessary not old, or is the piglit test too strict for bri-linear?

> The question is whether you really want to do something
> like that for explicit lod, where the user supposedly exactly knows
> what he wants.  "textureLod" is not used often at all[1], so one can
> think that when it's used you'd better do it precisely.

Maybe, but I'd rather not remove an optimization for the sake of a test. If it gets to that point, then I'd prefer make brilinear optional based on mipmaps being properly constructured mipfilters or not.  Either way, I wanna fully understand what's going wrong before deciding.

Jose