^eMpTy^ said:Anyone have any links to anything showing that shader replacement hurt image quality in any way?
Shader replacements in a game helps the consumer since the game will perform better. An IHV replacing some math/shaders with something that does less work just to inflate their scores in a synthethic benchmark is something different since there's no benefit for the consumer. Games are of course used as benchmarks but a syntethic benchmark are supposed to show a more general performance indication. A game benchmark shows only the performance in that game.
Then there's of course the problem with "some shaders run better on hardware A, some on hardware B, and if hardware B can do D more efficient by doing E instead, without any IQ problems, why shouldn't it be allowed to do that?". The next 3D Mark (coming in 2 weeks according to the Inq) should fix these problems by using dynamically built shaders. So the NV40 will have one path, the R420 another ... if i'm not mistaken.