I added per GLRenderer::render traces - mir has to render each of the paused application surfaces one by one with the phone shell on top. Each of them take between 1.8 and 2.8 ms just within the render function... depending on other interactions we might miss the 16 ms after having four suspended applications (== 5 nearly fullscreen surfaces to blend)
I added per GLRenderer::render traces - mir has to render each of the paused application surfaces one by one with the phone shell on top. Each of them take between 1.8 and 2.8 ms just within the render function... depending on other interactions we might miss the 16 ms after having four suspended applications (== 5 nearly fullscreen surfaces to blend)