Your VM will work much faster if you take the time to do a second pass over the generated "code" and resolve the jumps to LABELs all at once instead of linearly scanning the code for the appropriate target on each jump.