without subgrid scale model was better because there was already enough numerical dissipation so no subgrid scale dissipation was necessary