Niklas KorzThere currently are two recent #Phoronix articles that are shared widely, comparing #KDE KWin 6.6 vs #Gnome Mutter 50 in terms of game performance (one for AMD, one for Nvidia). What it lacks to me is a disclosure whether this is comparing the pure #Wayland compositor performance or more specifically their #Xwayland integrations. At least for the Windows games I will have to assume it is the latter (as Proton does not yet _default_ to Wayland).
Now for gaming, Xwayland is still very much relevant, but considering that features like HDR are already exclusive to Proton Wayland, at least showing a comparison of this mode and removing Xwayland from the equation is a requirement for me to take these benchmarks seriously.
Generally speaking, #Xwayland should be avoided for gaming where possible (which still isn't always possible, especially until Steam fix their overlay). The high latency people experience after switching from #X11 to #Wayland does not come from Wayland itself, but the indirection of this legacy compatibility layer. Don't take my word for it, this was measured in detail by David Justo this year:
Xargos@niklaskorz Exactly I thought about it too. It would've been interesting too see stats with proton_enable_wayland on top of GE.