I'm still working on the subject, but I found that the tests are not good because the micro I use at 64Mhz internal clock when used from 16Mhz and consecutive pins of the same port are controlled (for example: PORTC.0 = 1, PORTC.1 = 0, etc) as outputs fail and do not stay at the designated value.
So I have ordered other micros to replace them, although there is no problem in using the old micro with lower clock speeds or adding a NOP after each PORTx.x and it works perfectly at any speed. Unfortunately I dismantled the display before knowing this and started with other things so I have postponed it for September. But I was also waiting for two suggestions for improvements to the compiler, one of which has already been incorporated and is that functions allow more than 5 input arguments, currently up to 30, and another of lesser importance and is that tables can be made with a greater number of elements, at least 255, which until now is limited to just over 100.