Benchmark test difference

Hello everyone,

I met the Lvgl library two months ago. During this time, the library really caught my attention. It was fun and enjoyable to spend time understanding it. I would like to thank the people who contributed to the creation of the library.

The equipment and ide I used are as follows,

Esp32 devkit module
Ili9341 TFT screen 320x240
XPT2046 resistive touch screen
Arduino ide - 2.3.4
Vs code - 1.97.0
Tft_espi by Bodmer - 2.5.43
XPT2064_Touchscreen by Poul Stoffregen - 1.4
Lvgl - 9.2.2

I did standard Lvgl Benchmark tests with Arduino ide and Vs code. During these tests, I saw that there were 10% FPS differences despite using the same libraries and the same main codes.

Does anyone have any thoughts on this?

Thanks…

I am not sure how you are using the same libraries in VS code as in the Arduino IDE. You can’t be. The ESP-IDF is not available for the Arduino IDE and what runs for an SDK for the ESP32 in the Arduino IDE is not going to work in anything but the Arduino IDE.

I compiled the main code with arduino ide and got the benchmark results. Then I compiled the same main code with VS code and compared the benchmark results. The code compiled by arduino ide produced 10% better FPS.

It could be from the optimization flag the compiler is getting. You need to double check what the compiler flags are. Look for a flag -Ox where x is a number. The higher the number the faster the code is going to run. The number can be 1 to 3.