Three comparison/branch pairs are not free.

Clang will optimize

  bool swap_if(bool c, 
      int& a, int& b) {
    int ta = a, tb = b;
    a = (-c & tb)|((c-1) & ta);
    b = (-c & ta)|((c-1) & tb);
    return c;
  }

into cmov instructions. Used in a quicksort partition loop as

  l += swap_if(
    *r <= pv, *l, *r);

It makes quicksort fully 2x as fast as the usual branch.