Optimize cache and memory access

Some optimized tricks for AVR are included in this commit.
I spent a lot of time optimizing every line in these functions, looking for instruction-level optimization using godbolt.
A large portion of the time was spent on GDB, since I forgot bunch of checks and variable changes when I optimize these function.
Don't be stupid like me

Fix missing result place-holder variable in the previous commit

Fix readme

Add Youtube video

Merge branch 'main' of https://github.com/raspiduino/arv32-opt

24/9/2023