TFLOPs generally correlate to actual, general game performance quite well.

A very demanding game will do poorly on a gpu with a given TFLOPs score, but a less demanding one may do fairly well.

So, if you can roughly dial things into a TFLOPs value, you can get a result that is then generally applicable to all games.

The most graphically demanding games require the most TFLOPs out of a GPU to render well and fast, but less demanding games can be fine with a lower TFLOPs GPU.

Like… a GPU with a medium high TFLOPs score may only be able to push about 50 fps on CP77 ultra w/o RT, but it could likely easily hit something like 200ish FPS in say, CounterStike2, both games at 4k.

https://www.resetera.com/threads/all-games-with-ps5-pro-enhancements.1026072/

A PS5Pro can render CP77 at ‘quality mode’ …which is roughly the ‘high’ graphical quality preset of CP77 on PCs, with… I believe all raytracing other than raytraced shadows off…

… at 2k30fps, or it can render at a dynamicly scaling resolution between 2k and 1080p at a locked 60fps.

https://www.gpu-monkey.com/en/benchmark-amd_radeon_rx_6700-cyberpunk_2077

An RX 6700 can render CP77 at 2k, 43 fps avg.

This site isn’t 100% clear as to whether these GPU scores are done at the ‘high’ preset, or the ‘ultra’, all settings maxed out preset… but with all raytracing off.

https://www.gpu-monkey.com/en/benchmark-amd_radeon_rx_6600-cyberpunk_2077

A 6600 can do the same, high or ultra preset, with no raytracing, at 2k, with 31 fps.

1080p at 51 fps.

So uh yeah, even with the top of the line PS5 variant, the PS5Pro’s graphical capabilities are a bit worse than a 6700, and a bit better than a 6600.

I have used many AMD cards, and yeah basically, if you only use raytraced local shadows?

At 1080p, or 2k?

Very roughly, you’ll get about the same fps with a ‘high’ preset, and only raytraced local shadows, as you would with no rt local shadows, but the game’s graphical preset bumped up to ‘ultra’.

There are now like 6 different raytracing options in the PC version of CP77.

They only work well with Nvidia cards… and the latest iteration of AMD cards, that now have basically their equivalent of RT cores, though they’re not as powerful, nor as expensive, as Nvidia cards.

Needless to say, PS5… normals, or whatever, are going to be even weaker than the ‘somewhere between an rx 6600 and rx 6700’ rough equivalence.

…

Playstation users only even see about 10% of all the graphical settings options that PC players see, really just the 4 post processing effects and motiin blur. PC players get many, many more settings they can tweak around.

Also, PS5 users generally seem to think CP77 is being rendered at 4k, in quality mode.

Uh, no, it isn’t, its rendered at 2k max, and that is likely using FSR 2.1,with a bit of dynamic frame upscaling within CP77, meaning the actual real render resolution is roughly 5% to 15 % lower than 2k, then FSR upscales to 2k,… then the PS5 does its own upscaling outside of the game, likely via an emulator style simple integer upscale algo, to display at 4k on a 4k TV.

… I am saying all of this with confidence because I have spent a considerable amount of time fucking with customizing dynamjc resolution scaling, and other graphical settings, in CP77, on different AMD cards, on PCs.

The PS5Pro is using the same dynamic resolution scaling that exists in CP77 on PCs. PCs generally do best with an ‘Auto’ mode for various frame upscalers, but the dynamic setting on PC basically lets you throw out the auto settings and fuck about with your own custom tolerances for max and min resolutions and frame rates.