Would directly using CPU rendering for 1-bit MLCD be faster?

Troubleshooting
1

The current development version being used is sf32lb52, and I noticed that LVGL’s GPU acceleration is optimized for multi-color rendering. If I’m actually driving a 1-bit MLCD display, I’d have to first render in multi-color and then downsample it to black-and-white monochrome.

I’m torn between whether it’s better to just use CPU rendering or go with GPU. The GPU is faster, but its performance is wasted on multi-color rendering :thinking:

2

You can use GPU rendering, and then convert the buffer to 1-bit when sending to the display.