Been playing more Godfall recently, it further reinforced my earlier impression with how much "sparkle" you lose with FSR. Settled on TAAU 50% as even better than FSR UQ 4K, but as I have noticed before the temporal stability of simple TAA (and thus for FSR) is much better than TAAU when panning through sideways movement. The card can do native 4K but been enjoying the lower power usage (chip hovers around 100w only, with no UV!) and it's become a pet interest of mine to try and run games at more efficient power targets. It's one aspect I wish there were more tests on, particularly for DLSS as well, I reckon there's a great opportunity for power savings through that considering how well it works when working with fewer pixels. I know Igor's done some with DLSS 1.0 but haven't seen much testing done on it since.
It's actually been incredible to see how power efficient the cards are especially when optimising for good PQ rather than just maxing out the sliders. In trying to do some testing for the steam deck it was actually funny to see just how well even Cyberpunk w/ RT (on RDNA2 no less!) can run on 50-60w of chip power. I wonder now if the MCM beasts that are to come won't have even more incredible ppw with some tweaks.
It's actually been incredible to see how power efficient the cards are especially when optimising for good PQ rather than just maxing out the sliders. In trying to do some testing for the steam deck it was actually funny to see just how well even Cyberpunk w/ RT (on RDNA2 no less!) can run on 50-60w of chip power. I wonder now if the MCM beasts that are to come won't have even more incredible ppw with some tweaks.