RE: [Algorithms] Terrain performance comparisons

[De B. <be...@ii...>]

Everyone should read the latest GD mag about fill rate bound and transform
bound geometry - "Stop Hitting the BOTTLENECK". Basically if you have large
walls you may have much more fill stuff happening then transform. Often both
can occur in parallel and so an object spends more time doing fill, then it
is fill bound.

If you have a large polygon that has 4 passes doing fancy bump mapping,
surely its the shader LOD that is more important since it will be fill
bound. Of course if all your geometry is highly tesselated and in one pass
then yes geometry LOD is more important and more "compelling". I recon in
general shader LOD is just as important then geometry LOD (of course in
particular scenes one may be more useful then the other).

>>Shader LOD is less compelling to me than geometry LOD. If you have
>>pretty much any sort of LOD - even static levels - then you're going to
>>be drawing (very) roughly constant-area triangles. So your
>>vertex-processing load per pixel is roughly constant. And obviously your
>>pixel load per pixel is constant (excepting overdraw, which you can't do
>>much about using LOD - you need occlusion algos, which is orthogonal to
>>the problem).

-Ryan De Boer