After reading the Tony Tamasi interview today I'm convinced that NVidia will position the NV3x against the R(V)3XX by saying that FP24 is inadequate and FP32 is essential for complex shaders (it's pretty much the only advantage they have left if the R350 supports longer-than-NV30 shaders). I'm basically putting this thread here to record my prediction.
This is the first time I've really heard NVidia talking as though FP24 is inadequate, and I think the enhanced shader capabilities in the R350 is the reason.
This is the first time I've really heard NVidia talking as though FP24 is inadequate, and I think the enhanced shader capabilities in the R350 is the reason.