Even if a lot of the memory use is in VRAM, it's an improvement. It means you have more RAM available for apps, instead of graphic buffers, and that those are where they should be processed (which means they will be processed faster, because you don't need to wait for the bus). Also, this memory use will not be probably from Wayland, but from the apps, which is not really relevant when comparing Wayland to X.org. Of course, if X.org does it the inefficient way, it will show as consuming more than it's actually relevant.
Originally Posted by doom_Oo7
It's not that hard to build for different architectures. Just make two different build directories and build for one target on each. IIRC (I'm more comfortable with CMake, since I just make some toolchains to adapt it to whatever I want) it's just running, from the build dir, ../configure --target=i686-gnu-linux or something like that. From that point, you pull, you make, you are happy.
Originally Posted by ethana2
If that's the case, it's an improvement already, at least for desktops.
Originally Posted by Ancurio