Not necessarily since I have a few machines that need to be on 24/7 but quite often they don't do too much during the early hours of the morning, so why waste power doing nothing when you could do something with those idle cycles to crunch some climate models.
Anyway look at it this way, those climate models are going to be run at some point anyway, so the energy required to do this will be used at some point.
It was worked out that the amount of energy used to make tea in the UK alone is 23 times that required to run the entire climate prediction experiment, and that assume 100% efficiency to boil the water.
Also most people running the Climateprediction models are more aware of their carbon footprints anyway and ofset this in other ways.
I think the benefits of crunching these models outweigh the cost in consumed energy.