I've been testing out beta and have noticed the Vulkan backend using around 30mb less memory than GL on Linux. Hopefully we'll see even more reductions when QML

I've been testing out beta and have noticed the Vulkan backend using around 30mb less memory than GL on Linux. Hopefully we'll see even more reductions when QML3 is released (which really focuses on embedded). Another interesting thing is that QML now has cross-platform desktop widget via QtQuickControls2 (i.e. native Windows and MacOS styles) and the improved CMake support is also pretty convenient. I can now get rid of my .qrc files and just write everything in CMake. I really admire how smooth this migration was and it shows Qt really cares about minimizing breaking changes.