.copy_from_slice() does the same job of .clone_from_slice(), but the
former is explicitly for Copy elements and calls `memcpy` directly, and
thus is it efficient without optimization too.