And it did that with 1/8th the budget.
That's probably another one of those "facts" you found via your magic ass, right?
There is no way in hell, Fiji's (inc the ancillary tech - interposer design, HBM R&D etc) R&D costs equate to an eighth of what GM204 cost to develop. Even a simple glance at the R&D spend for each company (
AMD and
Nvidia) should tell you that is basically impossible.
If AMD can develop a 596mm² GPU utilizing a swath of new GPU tech for 1/8th the R&D budget of a 398mm² GPU based on refinement of previous design, it makes you wonder why they didn't apply the same fervour to some much smaller chips.
What an arrogant comment! Least this card can count to 4...
Well played!
What is "performance density"?
Probably someone in marketing noticed they didn't have enough bullet points on the slide then got the idea to divide transistor count per mm by some hand-tailored game image quality level and viola "Performance Density"!!
If AMD lead this particular metric, it should become the new "must have" feature very soon. Hallelujah! I was getting sick of DX12 feature level support being the defining factor for buying anything graphics related.