Gathering multiple system parameters in a single call

I should note that I think the schedulers could use some fine-tuning on this, actually: I understand there is a latency (and accompanying power consumption) going from (say) C10 to C0, so it'd make sense to completely stop using some cores, but that isn't what happens, based on what I see in Powertop: every core sees some use. Of course, I don't actually know what the most efficent thing is: just an observation from my machine.