When exactly does the Geforce 6800 NDA embargo lift?

[AIWGuru]

also understand the concept of different architectures you guys are trying to explain in THEORY U ARE RIGHT. See this pipeline stuff works for CPUs, but i have yet to see a real world exapmle where a gpu's shorter pipeline or archicture makes such a dramatic difference. I am not saying it's impossible for a slower clocked gpu to be faster in theory, but in real world, unlike the cpu compeittion between intel and amd, the graphics industry is much closer in design that is why it's more competitive because every ounce of speed counts.

Well then you're blind.
When Nvidia created the NV30 architecture they went with VLIW which is signifigantly less efficient per clock than anything else out there. This is why they've reverted. This efficiency is reflected in operations which are no reliant on memory bandwidth. I'm read what you've said *carefully* and come to the conclusion that you really don't know.
You do make a semi valid point about GPUs being less affected by long pipes and instruction sets than CPUs since most of their work used to be SIMD. Now, with shaders GPUs are basically CPUs and the penalty is just as high.

 

[AIWGuru]

" Since we "only" have 16 texture units coming out of the pipe, on older games it will be hard to get more than 2x performance per clock with NV40, but for newer games with single textured and pixel shaded rendering, we could see anywhere from 4x to 8x performance gain per clock cycle when compared to NV3x. " - AT Review
-HA HA
-told you so
-neener neener neener
-so on