pascal is just maxwell at higher clocks

I agree with this guy's findings, interesting video here. This in my opinion is what professional reviewers should do instead of just running dumb benchmarks and recording results.

https://youtu.be/nDaekpMBYUA?t=448

He throttled the 1080 to get clocks comparable to the 980ti, then made some calculations to adjust for shader count and he posted the results in the video. The 980ti was actually slightly quicker when the 1080 lost its clock speed advantage.

Of course its not quite as simple as this as nvidia managed to boost these clocks whilst reducing voltage which in turn reduced heat and power usage. But the clock for clock performance is more or less equal.

That doesn't seem entirely surprising to me. It is a bit like saying if you took the turbos off of the engine on my car then it would perform like a car with the same engine without turbos...

A lot of the new chip generation design is in making it possible to run them at higher clock speeds. There are probably also new instructions in the chip but they will only make any difference once tests/applications/games are written to take advantage of the new instructions - running the same code on both chips probably would give similar results at similar clock speeds - running optimised code on the new chip may see a significantly different result.

Yes totally agree here

its just a tweaked a maxwell with 16nm

volta will be a completely new architecture from the ground up.

indeed, usually nvidia release 2 sets fo cards per tech, I think they pulled a bit of a fast one here, with the pascal name when it looks like its a maxwell refresh