Wasted 15mins of my life learning about minor optimizations in inline assembly, entire thing got btfo when I found out that Rust compiler doesn't give a fuck and picks optimal register size for an operation by default.
e.g. mov rax, 1 is actually mov eax, 1 in the final binary and me writing mov eax, 1 explicitly does nothing. Assemblyniggers of course lied to me by claiming that compilers are le bad or something...