Spot on. Why not use any more or less standard cpu benchmark? The publication is either being strangely incompetent at this, or deliberately misleading.