4 ms·

mov ax, 13h int 10h

• exDM69 14y ago
  You are correct. And using a mov to ax rather than xor+mov to al is counter-intuitively one byte smaller (there's one more byte of immediate value in the former). Every byte counts! (I tested using "nasm -f bin")
  • gliptic 14y ago
    However, in .coms you could assume that AX was 0 upon entry and skip the XOR.