I once made a blog about how assembly is bloated so today I decided to try fasm, it was amazing, it's almost as efficient as C generated ELF,
For example, using NASM (or YASM but the difference is only 0.1 KB if not less) a Hello world program would look like this:
BITS 64
segment .text
global _start
_start:
mov rax, 1
mov rdi, 1
mov rsi, m
mov rdx, ml
syscall
mov rax, 60
mov rdi, 0
syscall
segment .rodata
m: db "Hello world!", 10
ml: equ $ - m
And when compiled using:
$ nasm -felf64 a.asm && ld -o a a.o
Where a.asm is the assembly source code you see above you get a 8.7 KB binary
So now let's do the same but using FASM:
format ELF64 executable 3
segment readable executable
_start:
mov rax, 1
mov rdi, 1
mov rsi, m
mov rdx, ml
syscall
mov rax, 60
mov rdi, 0
syscall
segment readable
m: db "Hello world!", 10
ml = $ - m
The code hasn't changed much but when we compile this code using:
$ fasm a.asm && chmod a+rx ./a
Where a.asm is the assembly source code you see above you get a 235 B binary
That's literally 8.465 KB improvement for only changing 5 lines of code... That's only one byte larger than out source code -- 234 B
Crazy how fast, small and nice this assembler is, give it a try! :)