@Downpour@Natanox x11 allows 2d acceleration (when it implemented by ddx driver). In that case system will work faster. This works with intel (pre-skylake) and nouveau drivers. And most embedded-targeted drivers.
Usual configuration (modesetting + glamor) or with x11 compositing enabled usually works better with wayland. Wayland does not have 2d acceleration drivers (and almost does not have drawing which can be 2d accelerated) and you stick on opengl-based acceleration.
Only profit i getting in x11 over wayland+xwayland on amdgpu now is my custom dri3 implementation, allowing handling gpu reset without restarting Xorg (have not publiushed it yet, but it’s based on and very close to tigerVNC’s dri3). But even on wayland, x11 apps may use less resources than wayland-native, because x11 drawing apis availiable and apps do not need load libGL+llvm+other stuff to draw UI, only need load libX11/xcb and send drawing commands to Xorg/Xwayland…
@Downpour @Natanox x11 allows 2d acceleration (when it implemented by ddx driver). In that case system will work faster. This works with intel (pre-skylake) and nouveau drivers. And most embedded-targeted drivers.
Usual configuration (modesetting + glamor) or with x11 compositing enabled usually works better with wayland. Wayland does not have 2d acceleration drivers (and almost does not have drawing which can be 2d accelerated) and you stick on opengl-based acceleration.
Only profit i getting in x11 over wayland+xwayland on amdgpu now is my custom dri3 implementation, allowing handling gpu reset without restarting Xorg (have not publiushed it yet, but it’s based on and very close to tigerVNC’s dri3). But even on wayland, x11 apps may use less resources than wayland-native, because x11 drawing apis availiable and apps do not need load libGL+llvm+other stuff to draw UI, only need load libX11/xcb and send drawing commands to Xorg/Xwayland…