We did a trial where we looked at how much energy we used on a weekend where the computers were not switched off at the wall compared to a weekend where they were. We made significant reductions in our baseload energy usage when we switched them off at the wall each weekend.