https://bugs.kde.org/show_bug.cgi?id=477016
--- Comment #32 from Zamundaaa <xaver.h...@gmail.com> --- > wouldn't VRR Off just be pushing X frames per second, resulting in the same > amount of GPU usage as a VRR on pushing the same X frames per second? On a very basic setup, yes. When the driver knows for certain that the content on the screen doesn't change though, it can optimize a lot - for example on laptops, tech like PSR allow it to just tell the display that it should keep showing the last frame, without needing to keep sending the same image to the display all the time. There is a way to tell the driver which parts of the image changed (and I think you can send an empty rectangle too) but I'll have to experiment with that to figure out how well it works / how much the drivers actually use that. There's also CPU usage; telling the kernel to re-submit the last image takes some amount of CPU power and wakes the CPU up frequently, while the GPU passively sending the last image repeatedly doesn't require the CPU to do anything. -- You are receiving this mail because: You are watching all bug changes.
