Hi,
As we’re searching for the right platform to run LVGL on, we discovered a weird scenario that left me scratching my head a bit.
The first platform we tried was an STM32F748 with external ram and flash, running at 480MHz. With that, we got 93FPS in the benchmark.
Since this architecture didn’t prove enough for our PoC we went to an IMX6ULL module with Linux (Colibri toradex module). It’s a single core 900 MHz processor. The odd thing is that the benchmark performance of this setup with framebuffer output was only 27FPS, well below what we got with a microcontroller that’s only half its frequency. There’s also an LVGL port on GitHub done on an IMX93 that had similar results, so while our quick test definite didn’t get all the performance of the module, I don’t think we were too far off.
Any ideas on why got such different results than what you’d normally expect?