The library had quite a bit of overhead from the set_pixel calls. This uses the faster method of specifying an area and then sending all the pixels in sequence.
Some benchmarks show a speed increase of ~20 times on an stm32f103
I'm not sure what a good buffer size is. 32 seems to give a nice speed improvement, but I imagine it becoming slightly faster with higher buffer sizes, at the cost of more RAM of course
sajattack
commented
4 years ago
The library had quite a bit of overhead from the set_pixel calls. This uses the faster method of specifying an area and then sending all the pixels in sequence.
Some benchmarks show a speed increase of ~20 times on an stm32f103
I'm not sure what a good buffer size is. 32 seems to give a nice speed improvement, but I imagine it becoming slightly faster with higher buffer sizes, at the cost of more RAM of course