Getting better all the time!
I’m curious if that fixes the issue I was facing. FSR2 (and 1 for that matter) basically did nothing for performance, or even reduces it when enabled. FSR3 on the other hand often gives me like up to 40 additional FPS in some games, but of course not all games have FSR3.
I still see the same old issue of a lot of Proton games degrading in FPS when you change graphic settings though, requiring a restart of the game to properly test and optimize them. No idea if that’s an AMD or Proton specific issue though.
This is bad reporting from phoronix (not surprising). The performance bug has nothing to do with FSR. It was just discovered in an FSR demo.
The bug was in radv which affected fsr compare to amdpro what did they get wrong in reporting?
No, just the FSR2 demo application, which suffered from poor performance regardless of whether or not FSR2 is on.
It is always these darn valve engineers! Why are they so good?!
If Valve’s Employee Handbook is to be believed, they don’t use a formal project structure with static teams. Instead each developer works on whatever project interests them, and one of Valve’s current goals is to improve game performance on Linux/AMD by contributing to upstream open source projects.
Valve is as close as we’ve gotten to someone paying a bunch of industry veterans to contribute to open source. It’s amazing what happens when all innovation isn’t black-boxed in an internal repository and forgotten about.
Valve directly profiting from all those games being available on their platform. Them helping out is a win-win-win situation.