i believe this was about benchmarks, not doom3. i found this to be an excellent article that matches my views of the benchmarking industry almost to the point. and doom3 as a benchmark? its not an equal playground either, whith all the pro nv30 optimations carmack keeps talking about.

"Futuremark = synthetic benchmark!!!! Synthetic benchmark = non real world results!!!! Non real world results = nothing like a real game!!!" so its just a coincidence that the FX has been displaying poor PS performance in TR, HL2 and Halo? benchmarks have their place, to make make an unbiased testing environment, and test tomorrows technologies on todays cards. these things have been severely compromised. and i dont think nvidia cares about what people think about 3dmark or benchmarks as a whole. As long as OEM manufacturers use them to select the cards that go into their pc configurations, nvidia will be onserned about it.

nvidia's fp32 may be slow, but it's far superior to ATI's atm.. ATI's current parts do fp24 and fp24 only :op

remind anyone of the old 3dfx v's nvidia 'you dont need 32bpp you can't tell the difference and we're faster" thing ? pro market needs fp32 and the pro nvidia boards are wiping the floor with ATI, and hell carmaks already been talking about fp64